【#1】World Economic Outlook Update, January 2021: Tentative Stabilization, Sluggish Recovery?

Tentative Stabilization, Sluggish Recovery?

  • Global growth is projected to rise from an estimated 2.9 percent in 2021 to 3.3 percent in 2021 and 3.4 percent for 2021-a downward revision of 0.1 percentage point for 2021 and 2021 and 0.2 for 2021 compared to those in the October World Economic Outlook (WEO). The downward revision primarily reflects negative surprises to economic activity in a few emerging market economies, notably India, which led to a reassessment of growth prospects over the next two years. In a few cases, this reassessment also reflects the impact of increased social unrest.
  • On the positive side, market sentiment has been boosted by tentative signs that manufacturing activity and global trade are bottoming out, a broad-based shift toward accommodative monetary policy, intermittent favorable news on US-China trade negotiations, and diminished fears of a no-deal Brexit, leading to some retreat from the risk-off environment that had set in at the time of the October WEO. However, few signs of turning points are yet visible in global macroeconomic data.
  • While the baseline growth projection is weaker, developments since the fall of 2021 point to a set of risks to global activity that is less tilted to the downside compared to the October 2021 WEO. These early signs of stabilization could persist and eventually reinforce the link between still-resilient consumer spending and improved business spending. Additional support could come from fading idiosyncratic drags in key emerging markets coupled with the effects of monetary easing. Downside risks, however, remain prominent, including rising geopolitical tensions, notably between the United States and Iran, intensifying social unrest, further worsening of relations between the United States and its trading partners, and deepening economic frictions between other countries. A materialization of these risks could lead to rapidly deteriorating sentiment, causing global growth to fall below the projected baseline.
  • Stronger multilateral cooperation and a more balanced policy mix at the national level, considering available monetary and fiscal space, are essential for strengthening economic activity and forestalling downside risks. Building financial resilience, strengthening growth potential, and enhancing inclusiveness remain overarching goals. Closer cross-border cooperation is needed in multiple areas, to address grievances with the rules-based trading system, curb greenhouse gas emissions, and strengthen the international tax architecture. National-level policies should provide timely demand support as needed, using both fiscal and monetary levers depending on available policy room.

Recent Developments and Implications for the Forecast

Trade policy uncertainty, geopolitical tensions, and idiosyncratic stress in key emerging market economies continued to weigh on global economic activity-especially manufacturing and trade-in the second half of 2021. Intensifying social unrest in several countries posed new challenges, as did weather-related disasters-from hurricanes in the Caribbean, to drought and bushfires in Australia, floods in eastern Africa, and drought in southern Africa.

Despite these headwinds, some indications emerged toward year-end that global growth may be bottoming out. Moreover, monetary policy easing continued into the second half of 2021 in several economies. Adding to the substantial support the easing provided earlier in 2021, its lagged effects should help global activity recover in early 2021. As discussed below, the 2021 global growth estimate and 2021 projection would have been 0.5 percentage point lower in each year without monetary stimulus.

High frequency indicators for the fourth quarter tentatively suggest momentum stabilized at a sluggish pace, helped by the broad-based shift earlier in the year toward accommodative monetary policy and fiscal easing in some countries (including China, Korea, and the United States). Temporary factors that had slowed global manufacturing-auto sector adjustments to new emissions standards, a lull in the launch of new tech products, and inventory accumulation-appeared to fade. Business sentiment and the outlook of purchase managers in the manufacturing sector ceased deteriorating, but remained pessimistic overall. Importantly, the new orders subcomponent of the surveys picked up, particularly in emerging market economies. Consistent with the surveys, world trade growth appeared to be bottoming out. Service sector activity on the other hand weakened somewhat but remained in expansionary territory, supported by still-resilient consumer spending-which, in turn, helped maintain tight labor markets, low unemployment, and modestly rising wages.

The main considerations for the global growth forecast from the backdrop of recent developments include: carryover from weaker-than anticipated second half outturns for 2021 among key emerging market economies; signs of tentative stabilization in manufacturing in the fourth quarter, but some weakening in still-resilient service sector activity; accommodative financial conditions; and uncertain prospects regarding tariffs, social unrest, and geopolitical tensions.

Global Growth Outlook: Modest Pickup in 2021

Global growth, estimated at 2.9 percent in 2021, is projected to increase to 3.3 percent in 2021 and inch up further to 3.4 percent in 2021. Compared to the October WEO forecast, the estimate for 2021 and the projection for 2021 repsent 0.1 percentage point reductions for each year while that for 2021 is 0.2 percentage point lower. A more subdued growth forecast for India (discussed below) accounts for the lion’s share of the downward revisions.

These outcomes depend to an important extent on avoiding further escalation in the US-China trade tensions (and, more broadly, on pventing a further worsening of US-China economic relations, including around tech supply chains), averting a no-deal Brexit, and the economic ramifications of social unrest and geopolitical tensions remaining contained.

  • In the United States, growth is expected to moderate from 2.3 percent in 2021 to 2 percent in 2021 and decline further to 1.7 percent in 2021 (0.1 percentage point lower for 2021 compared to the October WEO). The moderation reflects a return to a neutral fiscal stance and anticipated waning support from further loosening of financial conditions.
  • Growth in the euro area is projected to pick up from 1.2 percent in 2021 to 1.3 percent in 2021 (a downward revision of 0.1 percentage point) and 1.4 percent in 2021. Projected improvements in external demand support the anticipated firming of growth. The October 2021 WEO projections for France and Italy remain unchanged, but the projections have been marked down for 2021 in Germany, where manufacturing activity remains in contractionary territory in late 2021, and for Spain due to carryover from stronger-than-expected deceleration in domestic demand and exports in 2021.
  • In the United Kingdom, growth is expected to stabilize at 1.4 percent in 2021 and firm up to 1.5 percent in 2021-unchanged from the October WEO. The growth forecast assumes an orderly exit from the European Union at the end of January followed by a gradual transition to a new economic relationship.
  • Japan’s growth rate is projected to moderate from an estimated 1 percent in 2021 to 0.7 percent in 2021 (0.1 and 0.2 percentage point higher than in the October WEO). The upward revision to estimated 2021 growth reflects healthy private consumption, supported in part by government countermeasures that accompanied the October increase in the consumption tax rate, robust capital expenditure, and historical revisions to national accounts. The upgrade to the 2021 growth forecast reflects the anticipated boost from the December 2021 stimulus measures. Growth is expected to decline to 0.5 percent (close to potential) in 2021, as the impact of fiscal stimulus fades.

For the emerging market and developing economy group, growth is expected to increase to 4.4 percent in 2021 and 4.6 percent in 2021 (0.2 percentage point lower for both years than in the October WEO) from an estimated 3.7 percent in 2021. The growth profile for the group reflects a combination of projected recovery from deep downturns for stressed and underperforming emerging market economies and an ongoing structural slowdown in China.

  • Growth in emerging and developing Asia is forecast to inch up slightly from 5.6 percent in 2021 to 5.8 percent in 2021 and 5.9 percent in 2021 (0.2 and 0.3 percentage point lower for 2021 and 2021 compared to the October WEO). The growth markdown largely reflects a downward revision to India’s projection, where domestic demand has slowed more sharply than expected amid stress in the nonbank financial sector and a decline in credit growth. India‘s growth is estimated at 4.8 percent in 2021, projected to improve to 5.8 percent in 2021 and 6.5 percent in 2021 (1.2 and 0.9 percentage point lower than in the October WEO), supported by monetary and fiscal stimulus as well as subdued oil prices. Growth in China is projected to inch down from an estimated 6.1 percent in 2021 to 6.0 percent in 2021 and 5.8 percent in 2021. The envisaged partial rollback of past tariffs and pause in additional tariff hikes as part of a “Phase One” trade deal with the United States is likely to alleviate near-term cyclical weakness, resulting in a 0.2 percentage point upgrade to China’s 2021 growth forecast relative to the October WEO. However, unresolved disputes on broader US-China economic relations as well as needed domestic financial regulatory strengthening are expected to continue weighing on activity. After slowing to 4.7 percent in 2021, growth in ASEAN-5 countries is projected to remain stable in 2021 before picking up in 2021. Growth prospects have been revised down slightly for Indonesia and Thailand, where continued weakness in exports is also weighing on domestic demand.
  • Growth in emerging and developing Europe is expected to strengthen to around 2.5 percent in 2021-21 from 1.8 percent in 2021 (0.1 percentage point higher for 2021 than in the October WEO). The improvement reflects continued robust growth in central and eastern Europe, a pickup in activity in Russia, and ongoing recovery in Turkey as financing conditions turn less restrictive.
  • In Latin America, growth is projected to recover from an estimated 0.1 percent in 2021 to 1.6 percent in 2021 and 2.3 percent in 2021 (0.2 and 0.1 percentage point weaker respectively than in the October WEO). The revisions are due to a downgrade to Mexico‘s growth prospects in 2021-21, including due to continued weak investment, as well as a sizable markdown in the growth forecast for Chile, affected by social unrest. These revisions are partially offset by an upward revision to the 2021 forecast for Brazil, owing to improved sentiment following the passage of pension reform and the fading of supply disruptions in the mining sector.
  • Growth in the Middle East and Central Asia region is expected at 2.8 percent in 2021 (0.1 percentage point lower than in the October WEO), firming up to 3.2 percent in 2021. The downgrade for 2021 mostly reflects a downward revision to Saudi Arabia’s projection on expected weaker oil output growth following the OPEC+ decision in December to extend supply cuts. Prospects for several economies remain subdued owing to rising geopolitical tensions (Iran), social unrest (including in Iraq and Lebanon), and civil strife (Libya, Syria, Yemen).
  • In sub-Saharan Africa, growth is expected to strengthen to 3.5 percent in 2021-21 (from 3.3 percent in 2021). The projection is 0.1 percentage point lower than in the October WEO for 2021 and 0.2 percentage point weaker for 2021. This reflects downward revisions for South Africa (where structural constraints and deteriorating public finances are holding back business confidence and private investment) and for Ethiopia (where public sector consolidation, needed to contain debt vulnerabilities, is expected to weigh on growth).

Nonetheless, downside risks remain prominent.

  • Rising geopolitical tensions, notably between the United States and Iran, could disrupt global oil supply, hurt sentiment, and weaken already tentative business investment. Moreover, intensifying social unrest across many countries-reflecting, in some cases, the erosion of trust in established institutions and lack of repsentation in governance structures-could disrupt activity, complicate reform efforts and weaken sentiment, dragging growth lower than projected. Where these pssures compound ongoing deep slowdowns, for example among stressed and underperforming emerging market economies, the anticipated pickup in global growth-driven almost entirely by the projected improvement (in some cases, shallower contractions) for these economies-would fail to materialize.
  • Higher tariff barriers between the United States and its trading partners, notably China, have hurt business sentiment and compounded cyclical and structural slowdowns underway in many economies over the past year. The disputes have extended to technology, imperiling global supply chains. The rationale for protectionist acts has expanded to include national security or currency grounds. Prospects for a durable resolution to trade and technology tensions remain elusive, despite sporadic favorable news on ongoing negotiations. Further deterioration in economic relations between the United States and its trading partners (seen, for example, in frictions between the United States and the European Union), or in trade ties involving other countries, could undermine the nascent bottoming out of global manufacturing and trade, leading global growth to fall short of the baseline.
  • A materialization of any of these risks could trigger rapid shifts in financial sentiment, portfolio reallocations toward safe assets, and rising rollover risks for vulnerable corporate and sovereign borrowers. A widespad tightening of financial conditions would expose the financial vulnerabilities that have built up over years of low interest rates and further curtail spending on machinery, equipment, and household durables. The resulting renewed weakness in manufacturing could eventually spad to services and lead to a broader slowdown.
  • Weather-related disasters such as tropical storms, floods, heatwaves, droughts, and wildfires have imposed severe humanitarian costs and livelihood loss across multiple regions in recent years. Climate change, the driver of the increased frequency and intensity of weather-related disasters, already endangers health and economic outcomes, and not only in the directly affected regions. It could pose challenges to other areas that may not yet feel the direct effects, including by contributing to cross-border migration or financial stress (for instance, in the insurance sector). A continuation of the trends could inflict even bigger losses across more countries.

Policy Priorities

The risk of protracted subpar global growth remains tangible despite tentative signs of stabilizing momentum. Policy missteps at this stage would further enfeeble an already weak global economy. Instead, stronger multilateral cooperation and national-level policies that provide timely support could foster a sustained recovery to the benefit of all. Across all economies, a key imperative-increasingly relevant at a time of widening unrest-is to enhance inclusiveness, ensure that safety nets are indeed protecting the vulnerable, and governance structures strengthen social cohesion.

Multilateral cooperation. Closer cross-border cooperation is needed on multiple fronts. Countries should expeditiously address grievances with the rules-based trading system, promptly resolve the impasse over the World Trade Organization’s Appellate Body, and settle disagreements without raising tariffs and non-tariff barriers. Technology tensions between countries are unlikely to be resolved unless they cooperate to curtail cross-border cyberattacks and solve outstanding issues concerning intellectual property rights and technology transfer. Failure to work out trade and technology conflicts will undermine confidence further, weaken investment, and lead to rising job losses; over a longer horizon, this would inhibit productivity growth and slow increases in living standards. Countries urgently need to cooperate on curbing greenhouse gas emissions and limiting the rise of global temperature with an approach that ensures appropriate burden-sharing across and within borders. Other areas where stronger cooperation would help enhance inclusiveness and resilience include reducing cross-border tax evasion and corruption; avoiding a rollback of global financial regulatory reforms; and ensuring an adequately resourced global financial safety net.

Table 1. Overview of the World Economic Outlook Projections

(Percent change, unless noted otherwise)

Note: Real effective exchange rates are assumed to remain constant at the levels pvailing during October 14-November 11, 2021. Economies are listed on the basis of economic size. The aggregated quarterly data are seasonally adjusted. WEO = World Economic Outlook.

1/ Difference based on rounded ps for the current and October 2021 World Economic Outlook forecasts. Countries whose forecasts have been updated relative to October 2021 World Economic Outlook forecasts account for 90 percent of world GDP measured at purchasing-power-parity weights.

2/ For World Output, the quarterly estimates and projections account for approximately 90 percent of annual world output at purchasing-power-parity weights. For Emerging Market and Developing Economies, the quarterly estimates and projections account for approximately 80 percent of annual emerging market and developing economies’ output at purchasing-power-parity weights.

3/ Excludes the Group of Seven (Canada, France, Germany, Italy, Japan, United Kingdom, United States) and euro area countries.

4/ For India, data and forecasts are psented on a fiscal year basis and GDP from 2011 onward is based on GDP at market prices with fiscal year 2011/12 as a base year.

5/ Indonesia, Malaysia, Philippines, Thailand, Vietnam.

6/ Simple average of growth rates for export and import volumes (goods and services).

7/ Simple average of prices of UK Brent, Dubai Fateh, and West Texas Intermediate crude oil. The average price of oil in US dollars a barrel was $60.62 in 2021; the assumed price, based on futures markets (as of November 12, 2021), is $58.03 in 2021, and $55.31 in 2021.

8/ The annual average CPI for the United States is 2.3 percent in 2021 and 2.4 percent in 2021, the euro area is 1.4 percent in 2021 and 2021, and Japan is 1.1 percent in 2021 and 1.2 percent in 2021.

9/ Excludes Venezuela.

Box 1. Global Financial Conditions: Still Accommodative

Global financial conditions continue to be accommodative by historical standards. Overall conditions are little changed since the October 2021 Global Financial Stability Report, though there has been an easing in conditions in some inpidual economies (Figure 1).

Over the past three months, markets have again been driven by two main factors: monetary policy and investor perceptions about trade tensions. Monetary policy has remained supportive. For example, the US Federal Reserve cut its policy rate by 25 basis points; the European Central Bank restarted net asset purchases at a pace of €20 billion per month; the People’s Bank of China reduced its medium-term lending facility rate by 5 basis points; in Turkey, the central bank cut its policy rate by 450 basis points; while the central banks in Russia and Brazil reduced their interest rates by 75 and 100 basis points, respectively.

On trade tensions, the market has oscillated back and forth according to the latest trade-related news, including the recent announcement of a “Phase One” agreement on trade between the United States and China. On net, world equity markets have risen by about 8 percent over the past three months and long-term yields in the euro area, Japan and United States have increased by 15-30 basis points from very low levels.

【#2】Sơ Lược Về View Và Materialized View Trong Sql

Chào các bạn! Hôm nay của mình xin được chia sẻ về View và Materialized View. Trong khuôn khổ bài viết này, mình sử dụng PostgreSQL để thực hiện các ví dụ demo cho các bạn dễ hiểu.

OK. Giả sử ta có 1 yêu cầu thống kê điểm số trung bình mà người dùng xếp hạng cho mỗi sản phẩm. Ta sẽ cần dùng câu query như sau

Và kết quả mình thu được với dữ liệu trên như sau

Câu hỏi đặt ra ở đây là, thế mỗi lần ta muốn xem thống kê điểm số trung bình mà người dùng xếp hạng cho mỗi sản phẩm, ta lại phải đi gõ lại cái query dài ngoằng kia hay sao? Ở ví dụ của mình còn là 1 query khá đơn giản và chưa quá dài, nhưng cũng đủ ta phải ngán ngẩm khi gõ lại hoặc lôi nó vào trong code đúng ko nào?

PostgreSQL, MySQL, Oracle cho chúng ta 1 khái niệm “View” mà người ta hay gọi là “Khung nhìn” hay “Bảng tạm”. Chỉ cần google từ khóa “View trong SQL là gì?”, bạn có thể thấy vô vàng các kết quả. Một trong số đó tại http://viettuts.vn/sql/su-dung-view-trong-sql, họ đính nghĩa như sau

Có lẽ họ dịch google nên mới có chữ “Lượt xem” (views) như vậy. Tuy nhiên, phần nào đó cũng giúp các bạn hiểu sơ sơ View là gì. Đừng cố dịch View thành “khung nhìn”, nghe nó cứ điêu điêu thế nào ý. Tại https://www.postgresql.org/docs/9.2/static/sql-createview.html, họ định nghĩa rằng

CREATE VIEW defines a view of a query. The view is not physically materialized. Instead, the query is run every time the view is referenced in a query.

Bạn có thể hiểu một cách cực kỳ đơn giản, việc tạo 1 View có thể hiểu nôm na là tạo 1 Alias cho 1 câu query, tạo nên 1 “bảng tạm” . Khi bạn gọi đến view, đồng nghĩa với việc bạn gọi câu query mà bạn gán khi tạo VIEW. “Ở phần định nghĩa bên trên, họ nói rằng, VIEW thì không được lưu trữ vật lý, nó giống như một bảng ảo, Thay vào đó, câu query sẽ được chạy mỗi lần gọi tới VIEW.”

CREATE VIEW product_rate_statistic AS SELECT round(coalesce(avg(upr.score), 0), 2) AS avg_score, p.product_id, p.product_name, p.price, p.product_type FROM public.product p LEFT JOIN public.user_product_rate upr ON p.product_id = upr.product_id LEFT JOIN public."user" u ON upr.user_id = u.user_id GROUP BY p.product_id, p.product_name, p.price, p.product_type ORDER BY avg_score DESC;

Rất đơn giản phải không nào. Giờ ta test

Như vậy, việc gọi SELECT * FROM public.product_rate_statistic; tuơng đương với

Và sau khi insert INSERT INTO public.product(product_name, product_type, price) VALUES ('Product E', 'Draft', '10'); thì khi gọi lại View, ta được

Vì đặc điểm của View là không lưu trữ vật lý, do đó, mỗi lần gọi tới View, câu query đều được chạy lại. Đặc điểm đó cho ta 1 lưu thế, nghĩa là dữ liệu của View hiển thị là dữ liệu realtime (có thể mình dùng từ chưa được chuẩn), hay dữ liệu chính xác tại bất cứ thời điểm nào khi tham chiếu tới view.

Giả khả việc thống kê của bạn còn phụ thuộc tới hàng trăm nghìn các yếu tố khác, và mỗi lần query

mất tới vài giây. Vậy khi user load 1 sản phẩm để xem thông tin, user sẽ phải chờ đợi vài giây. Trong khi đó, yêu cầu hệ thống của bạn chỉ là : “Điểm của sản phẩm là điểm được thống kê và chốt từ quá khứ tớ thời điểm 24h ngày hôm trước”. Vậy thì ưu điểm của VIEW lúc này lại trở thành bất lợi khi mỗi lần gọi nó đều phải dựng lại kết quả trong khi điều đó không cần thiết. Lúc này, ta cần đến MATERIALIZED VIEW

MATERIALIZED VIEW cũng gần giống như VIEW, chỉ khác ở chỗ là nó được lưu trữ vật lý. Do đó khi bạn tham chiếu tới MATERIALIZED VIEW, CSDL của bạn sẽ không phải dựng lại kết quả của câu lệnh truy vấn ứng với view, mà chỉ là đi lôi KẾT QUẢ ĐANG LƯU TRỮ ra, giống như khi lôi 1 TABLE bình thường.

Vậy, KẾT QUẢ ĐANG LƯU TRỮ ở đây là gì? Nó sẽ là kết quả của câu truy vấn tại thời điểm tạo MATERIALIZED VIEW, hoặc kết quả của câu truy vấn tại thời điểm REFRESH MATERIALIZED VIEW gần nhất (hay cưới cùng tính tới thời điểm xét). Như vậy, với bài toán thống kê bên trên, ta sẽ refresh lại MATERIALIZED VIEW vào mỗi 24h hằng ngày là OK.

Cú pháp tạo


Cú pháp làm mới dữ liệu


CREATE MATERIALIZED VIEW public.product_rate_statistic_at_24h AS SELECT round(coalesce(avg(upr.score), 0), 2) AS avg_score, p.product_id, p.product_name, p.price, p.product_type FROM public.product p LEFT JOIN public.user_product_rate upr ON p.product_id = upr.product_id LEFT JOIN public."user" u ON upr.user_id = u.user_id GROUP BY p.product_id, p.product_name, p.price, p.product_type ORDER BY avg_score DESC;

Sau khi thêm dữ liệu product mới. INSERT INTO public.product(product_name, product_type, price) VALUES ('Product F', 'XXX', '10'); . Kết quả khi gọi MATERIALIZED VIEW không có gì thay đổi.

Đã có sự khác biệt đúng ko. Giờ các bạn đã hiểu rồi chứ.

https://www.postgresql.org/docs/9.2/static/sql-createview.html https://www.postgresql.org/docs/9.3/static/sql-creatematerializedview.html https://freetuts.net/database-view-la-gi-database-view-trong-mysql-239.html

All Rights Reserved

【#3】Materialize Các Phần Mềm Thay Thế Và Phần Mềm Tương Tự

Bootstrap là một bộ công cụ mã nguồn mở để tạo các trang web và ứng dụng trình duyệt. Nó chứa các mẫu thiết kế dựa trên HTML và CSS cho kiểu chữ, biểu mẫu, nút, biểu đồ, điều hướng và các thành phần giao diện khác, cũng như các phần mở rộng JavaScri…

Polymer là một thư viện sử dụng các công nghệ web mới nhất để cho phép bạn tạo các phần tử HTML tùy chỉnh. Xây dựng bất cứ thứ gì từ một nút đến một ứng dụng hoàn chỉnh như một yếu tố được đóng gói, có thể tái sử dụng, hoạt động trên máy tính để bàn…

Skeleton là một tập hợp nhỏ các tệp CSS & JS có thể giúp bạn nhanh chóng phát triển các trang web trông đẹp ở mọi kích cỡ, có thể là màn hình máy tính xách tay 17 “hoặc iPhone. Bộ xương được xây dựng trên ba nguyên tắc cốt lõi: – Lưới phản hồi x…

Metro UI CSS một tập hợp các kiểu để tạo một trang web có giao diện tương tự như Windows 8 .

Dự án Vật liệu góc là một triển khai của Thiết kế Vật liệu trong chúng tôi Dự án này cung cấp một tập hợp các thành phần UI có thể tái sử dụng, được kiểm tra tốt và có thể truy cập dựa trên hệ thống Thiết kế Vật liệu. Tương tự như bộ sưu tập các t…

Slides là một khung HTML và CSS đơn giản và mạnh mẽ cho phép bạn dễ dàng xây dựng các trang web tĩnh đẹp nhờ thiết lập mô-đun linh hoạt.

Flat UI Pro được tạo trên cơ sở Twitter Bootstrap theo phong cách phẳng tuyệt đẹp và bộ sản phẩm cũng bao gồm phiên bản PSD dành cho các nhà thiết kế. Flat UI Pro chứa một số lượng lớn các thành phần cơ bản, biểu tượng và glyphs.

GroundworkCSS là Khung HTML5, CSS và JavaScript đáp ứng miễn phí và nguồn mở 100%.

Hơn 400 thành phần UI vật liệu, hơn 600 biểu tượng vật liệu, 74 hình động CSS, tệp SASS, mẫu, hướng dẫn và nhiều hơn nữa. Miễn phí cho sử dụng cá nhân và thương mại. Được tin cậy bởi 250 000 + nhà phát triển và thiết kế. Được sử dụng bởi các công ty…

Spectre.css là một khung CSS nhẹ, đáp ứng và hiện đại để phát triển nhanh hơn và mở rộng.

Bộ công cụ UI chúng tôi 2.0 cho Web. https://github.com/ElemeFE/element

Responsable là một khung HTML LESS SCSS đáp ứng với đánh dấu sạch, CSS được chuẩn hóa và hình ảnh phản hồi. Nó cho phép các cột lồng nhau để dễ dàng lồng và tùy chỉnh lưới, và cung cấp các kiểu cơ sở cho các biểu mẫu, bảng và kiểu chữ để khởi động n…

Giải nén bữa ăn của bạn và nhận được mã hóa. Một thư viện CSS xâm lấn để bắt đầu phong cách của bạn.

【#4】Epigenetic Inheritance: Concepts, Mechanisms And Perspectives

  • Department of Physiology and Pharmacology, Sapienza University of Rome, Rome, Italy
  • Department of Psychology and “Daniel Bovet” Center, Sapienza University of Rome, Rome, Italy
  • Fondazione Santa Lucia, IRCCS, Rome, Italy

Parents’ stressful experiences can influence an offspring’s vulnerability to many pathological conditions, including psychopathologies, and their effects may even endure for several generations. Nevertheless, the cause of this phenomenon has not been determined, and only recently have scientists turned to epigenetics to answer this question. There is extensive literature on epigenetics, but no consensus exists with regard to how and what can (and must) be considered to study and define epigenetics processes and their inheritance. In this work, we aimed to clarify and systematize these concepts. To this end, we analyzed the dynamics of epigenetic changes over time in detail and defined three types of epigenetics: a direct form of epigenetics (DE) and two indirect epigenetic processes-within (WIE) and across (AIE). DE refers to changes that occur in the lifespan of an inpidual, due to direct experiences with his environment. WIE concerns changes that occur inside of the womb, due to events during gestation. Finally, AIE defines changes that affect the inpidual’s pdecessors (parents, grandparents, etc.), due to events that occur even long before conception and that are somehow (e.g., through gametes, the intrauterine environment setting) transmitted across generations. This distinction allows us to organize the main body of epigenetic evidence according to these categories and then focus on the latter (AIE), referring to it as a faster route of informational transmission across generations-compared with genetic inheritance-that guides human evolution in a Lamarckian (i.e., experience-dependent) manner. Of the molecular processes that are implicated in this phenomenon, well-known (methylation) and novel (non-coding RNA, ncRNA) regulatory mechanisms are converging. Our discussion of the chief methods that are used to study epigenetic inheritance highlights the most compelling technical and theoretical problems of this discipline. Experimental suggestions to expand this field are provided, and their practical and ethical implications are discussed extensively.


Many recent studies have demonstrated that stressful conditions that are experienced by parents can influence the offspring’s vulnerability to many pathological conditions, including psychopathologies-primarily related to a disruption in stress response mechanisms. These effects may even endure for several generations. Nevertheless, the mechanisms of this phenomenon have not been detailed, and only recently have scientists examined epigenetics to answer this question.

In this work, we systematize the concept of epigenetic inheritance, discuss the putative mechanisms, and recapitulate the methods for studying this circumstance, psenting their potentialities and limitations. We focus on the transmission of psychopathologies-especially in relation to disruptions in the stress response-because they have long been the center of the historical debate over the weights of genes and the environment in such processes as inpidual development and inheritance, given their complex nature and clear experience sensitivity. Thus, psychopathologies can be considered one of the most interesting and flourishing fields in the application of epigenetics.

Historical Background: Filling the Gap

Initially, and for a long time, parental influences on an offspring’s development were focused on two possible sources of variance: genes and the environment. Some scientists concentrated on how “slow and still” information could be transmitted to subsequent generations. The phylogenetic perspective was thus a central assumption, more or less implicit. The pmise was that genes themselves carry on blindly: the luckiest genes that are most well suited for the psent environmental conditions “win” and endure ( Dawkins, 1990).

This idea fits well with the Darwinian concept of adaptation as an all-or-nothing process, which can be recalled easily from the collective imaginary through such terms as “survival,” “reproductive power,” and “law of large numbers.” Those who survive live longer, thus theoretically increasing the probability of finding a mate and reproducing. This can be surely the case, but this theory alone is insufficient to explain phylogenetic development.

Yet, for a long time, genes and the environment were considered two separate aspects that interacted at the level of the phenotype. Even epigenetics was conceived of as being able to modify the genetic impact on an inpidual’s organization but remaining inside his existence (i.e., acting only during his lifespan). Until then, there was only one way in which the past could inform the coming new life: genes and parental care. Only when it was demonstrated that epigenetic modifications could be inherited did the ontogenetic (i.e., environmental influences) and phylogenetic (i.e., genetic determinants) worlds-which for years had approached each other in an asymptotic, exhaustive manner-finally merge at a new, theoretical intersection: epigenetic inheritance.

Epigenetics and Inheritance: Some Definitions

1. exposure to an event in generation F0.

2. an effect of the event must be observed in the third or fourth generation-i.e., F2 or F3-depending on whether the mother or father was first affected (F0).

Female exposure to a certain environmental factor during pgnancy might even affect the offspring’s germ cells directly, for which reason only the fourth generation can be considered “event-free” and unsullied. When a certain event produces an epigenetic change in the father, it can only modify his sperm, effecting reliable nongenetic inheritance in the third generation (Figure 1).

Figure 1. Transgenerational epigenetic inheritance. According to the classical definition of transgenerational epigenetic inheritance, environmental triggers that hit pgnant female inpiduals (F0) can affect “directly” not only the first new generation (F1), but also its germ cells that repsent the second generation (F2). For this reason, only changes in F3 can be due “purely” to epigenetic inheritance. The male germline, instead, can be affected only for one generation, allowing observing epigenetic inheritance already at F2.

This definition surely renders the observation of epigenetic inheritance easier, especially in humans, because it pvents the ambiguous interptation of data that are inevitably contaminated by other events that are not transmitted epigenetically through gamete programming. Nevertheless, this approach excludes the possibility of considering faster epigenetic effects, which are certainly more difficult to control experimentally but could still exist and have functions.

In fact, why must epigenetic transmission occur through germ cells and across several generations? The epigenetic modification of certain genes, produced by an environmental trigger, could lead to significant changes in an inpidual’s body that could persist over time and in turn signal the epigenetic reorganization of the subsequent generation. This phenomenon could happen without affecting the germline directly and despite the event that fostered such adaptation no being longer active once the embryo has begun its development. As we will see, experimental manipulation in animal models could overcome these problems. For this reason, we will attempt to unify and organize this potentially confusing terminological flowering in a coherent conceptual framework.

New Conceptualization

In recent years, many scientists have hypothesized and even demonstrated that certain experiences during the life of an inpidual influence the development of his offspring, even distally. It appears that some experiences modify genetic expssion, influencing:

1. how the organism itself responds to a changeable environment (i.e., more ontogenetic flexibility-direct or synchronous effect) and

2. how his descendants will increase their likelihood of surviving in a specific environment-that is, how information is transmitted to offspring regarding the environment that they will encounter (i.e., more phylogenetic flexibility-indirect and both synchronous and asynchronous effects).

Direct Epigenetics

The first aspect, which we will call direct epigenetics (DE), comprises all of the epigenetic changes that occur during an inpidual’s lifespan. Notably, this phenomenon implies even dynamic and short-term regulation of gene expssion, mediated by the action-almost in real time-of regulatory proteins, called transcription factors, such as c-fos, c-jun, ZENK and CREB. The genes that encode for such crucial functional elements are called immediate-early genes, because a change in their expssion is the first event that launches cascades of adaptive events, including the transcriptional aspects of other genes ( Johnson, 2010), ultimately producing even long-lasting effects.

Indirect Epigenetics

When an epigenetic change produced by a direct experience (DE) is transmitted to the offspring, that same experience becomes an indirect environmental trigger for the ontogenetic development of the new inpidual. Paralleling Crews (2008) and van Otterdijk and Michels (2016), the second form of environmental action (i.e., phylogenetic adaptation) can be pided into two categories of “indirect epigenetics (IE):” within and across. These two aspects can be considered the conceptual product of the historical development of this matter, the latter (across) being a more recent acquisition. Theoretically, these components are related and difficult to distinguish, even operationally.

Within indirect epigenetics (WIE) encompasses all of the epigenetic changes that act synchronously on the developing inpidual. Temporally, it starts at the very moment at which the zygote is formed and the environment begins changing. This category includes all of the factors that, more or less indirectly, can affect the developing inpidual, from the start to end of gestation. The underlying concept is that environmental changes occur when the (proto)-inpidual actually exists, synchronously.

Across indirect epigenetics (AIE) describes what happens from the moment of conception back toward the parents’ earlier life experiences (and even grandparents, as we will discuss), which asynchronously set the composition of germ cells (and possibly that of the intrauterine environment). Some authors have referred to all epigenetic changes that appear to be transmitted across generations as epimutations, in contrast to classical, less frequent genetic mutations ( Bennett-Baker et al., 2003). Notably, in this case, a certain event has consequences that are maintained over time, affecting the offspring’s destiny during gestation and, most importantly, later in life. Clearly, it is reasonable to believe that the closer we are to the moment of conception, the stronger the pdiction power of the variable is, or at least the easier it is to hypothesize a “causal” relationship, because it should be expected in an epistemology of complexity that conceives of development in terms of probabilistic epigenesis (see Gottlieb, 2007 for a theoretical detailed explanation). Nevertheless, as we will see, certain events can act as relevant pdictors even when distal in time.

We can discuss epigenetics only if a modification to gene expssion takes place. This idea, supported by Kovalchuk (2012), renders the function of the intrauterine environment in epigenetic transmission controversial-in cases in which environmental events produce changes that do not affect germ cells directly but persist and affect the newborn in later gestational stages. Nevertheless-and for this same reason-the epigenetic mechanisms that determine the womb cannot be neglected if they are demonstrated to mediate the transmission of information on the genetic expssion of the developing organism (as discussed below). As we will see, of all of the epigenetic mechanisms that are implied in these two indirect forms of transmission, maintenance methylation, de novo methylation and the regulatory and amplifying activities of ncRNA, are the most prominent. Although increasing data strongly suggest transgenerational inheritance of epigenetic information, the non-DNA-based processes by which information is transmitted across generations are largely unknown ( Houri-Zeevi and Rechavi, 2021).

Wider Clarifications and Considerations

The categorization above repsents a mere conceptual distinction that has been conceived simply to elucidate the phenomenon of interest. As a matter of fact, all of these aspects are expected to interact continuously, but we can distinguish, on a case-by-case basis, which conceptual element (ontogenetic vs. phylogenetic, direct vs. indirect, or even within vs. between) has more apparent relevance. Notably, our aim is to indicate that relegating epigenetic transmission only to the moment of gestation is impcise, hindering us from developing a wider and exhaustive understanding of this phenomenon. Moreover, our purpose is to merge and soften all dichotomic types of conceptualization, including those that we have proposed herein.

Wider environmental effectors, such as parental style and cultural aspects, must be considered with caution. They seem to be direct and indirect in their action, as well as synchronous and asynchronous. Certainly, they account for the general setting in which the newborn develops, which in turn begets different and complex forms of information about the past that guides ontogenetic and phylogenetic adaptation. This should be considered the most intuitive means of transgenerational transmission of information-the most naïve but still undeniable Lamarckian addition to Darwinian evolution. Nevertheless, and for this reason, the effect of these two variables is too complicated to account for, and studies that have attempted to demonstrate their function in epigenetic transmission (as we will see in the next section) are not exhaustive. Moreover, a discussion of these wider environmental factors is not pertinent to our discourse, given the level of inquiry that we are considering. Thus, we are setting aside these two aspects from our argumentation, except for pnatal maternal care and the few historical events that have been suggestive objects of study (e.g., the Dutch Famine and the Holocaust).

Epigenetic Spacetime

What does the concept of epigenetic inheritance add to science with respect to the earlier concept of evolution and genetic transmission? This type of communication appears to be faster and more contingent and thus more efficient. For this reason, epigenetics increases our heuristic power through a different concept of evolution, in which the environment has a more proactive role in influencing communication across generations, depending ultimately on two interconnected evolutionary processes: a Darwinian process (slow but steady) and a Lamarckian process (quick but labile). The historical contraposition of these processes is now evolving into a unified theory of evolution ( Skinner, 2021).

Therefore, we can imagine the transmission of information across generations as a succession of cycles, a sequence of light-cones that repsent the multidimensionality of a theoretical wave function that describes the amplitude of indetermination or the potential of the evolving system. Thus, we can picture the “pulsing” of this probabilistic informational mass unfolding across spacetime, merging at the moment of conception (considered our arbitrarily chosen observation point) and then expanding, only to collapse again (Figure 2).

Figure 2. Epigenetics through the Minkowskian cone. Epigenetic changes and related environmental factors visualized in 4D Minkowskian space, assuming conception as our arbitrarily chosen observation point, the zero of the system. Across indirect epigenetics (AIE) includes all those adaptations in parental life that pcede conception; within indirect epigenetics (WIE) describes all those changes that take place during the gestational period and, finally, direct epigenetics (DE) describes all those plastic processes that can occur after birth. Although these processes are strongly interconnected and can overlap on multiple levels in a complex real system, here they are treated as discrete and sequential, for the sake of clarity.

Evidence of Epigenetic Changes

There are several reviews on epigenetic changes ( Jawahar et al., 2021; Jung and Pfeifer, 2021; Conti and Alvares da Silva-Conforti, 2021; Maccari et al., 2021) and their heritability across generations ( Gapp et al., 2014; Skinner, 2014; Babenko et al., 2021; Bohacek and Mansuy, 2021; Szyf, 2021; van Otterdijk and Michels, 2021; Ambeskovic et al., 2021; Pang et al., 2021; Yeshurun and Hannan, 2021). We report only some of the most significant evidence to provide concrete examples of our proposed classification.

Evidence of Direct Epigenetics

Epigenetic changes can be triggered by several environmental factors, such as diet ( Mathers et al., 2010), pollution ( Christensen and Marsit, 2011), smoking ( Talikka et al., 2012), that can be labeled generically as “stressors,” referring to the neutral, adaptive meaning of the term ( Cabib and Puglisi-Allegra, 2012). Epigenetic aberrations have been implicated in many diseases, primarily cancer but also cardiovascular, autoimmune, metabolic and neurodegenerative diseases, often with particular regard to aging ( van Otterdijk et al., 2013; Jung and Pfeifer, 2021).

Animal Models

Several studies have examined the regulatory effects of adult stress on the methylation of the NR3C1 gene as a pathological marker and mediator of pathology, consequent to dysregulation of the hypothalamic-pituitary-adrenal axis (HPA), as exemplified in animal models of social defeat stress ( St-Cyr and McGowan, 2021). In addition, brain-derived neurotrophic factor (BDNF) is downregulated in several areas of the brain in animal models of depssion ( Elfving et al., 2010; Molteni et al., 2010; Qiao et al., 2014). Methylation of the promoter region of BDNF is associated with a reduction in hippocampal volume. In several animal models, hippocampal BDNF levels decline under acute ( Barrientos et al., 2003) and chronic ( Nibuya et al., 1995) stress conditions. Antidepssant treatment upgulates hippocampal BDNF, and knocking out BDNF in animal models impairs the response to treatment ( Khundakar and Zetterström, 2006; Monteggia et al., 2007).

Illustrating the unique functions of immediate-early genes (IEGs; cited above), CREB-mediated transient plasticity mediates the transition to an irreversible phenotype of addiction through the accumulation of delta-FosB, which mediates structural synaptic readaptations that strengthen themselves in a positive-feedback process ( Koob and Volkow, 2010).

The levels of certain stress-related proteins have been analyzed. CRFR1 levels are elevated in total KO (TKO) mice, as reported by Haramati et al. (2011). Similarly, Andolina et al. (2018) examined the effects of the absence of miR-34a/b/c (i.e., a TKO model, or TKO) on coping behavior. MiR-34 levels were assessed in all of the main areas that are involved in stress responses, peaking in DRN. TKO mice tended to resort to active coping behaviors when challenged by the forced swim test (FST). Accordingly, TKO mice overexpssed CRFR1 in DRN, compared with WT mice, and their “resilient” behavioral phenotype could be reverted through the injection of a CRFR1 antagonist, confirming its function in regulating stress response strategies.

Clinical Evidence

Evidence of Within Indirect Epigenetics

Animal Models

Animal models have been used to demonstrate the indirect effects of environmental factors during pgnancy on the offspring’s development. For example, maternal fat diet increases the susceptibility of male offspring to liver disease through epigenetic reprogramming of lipid metabolism and inflammatory responses ( Pruis et al., 2014). Further, pnatal undernutrition, for example, can permanently alter DNA methylation in the sperm of adult offspring in regions that are resistant to zygotic reprogramming, potentiating transgenerational transmission of metabolic disorders ( Radford et al., 2014).

Fetal programming also depends on miRNAs, although there is limited evidence in WIE. Placental miRNAs have been implicated by several groups ( Maccani et al., 2013; Morales-Prieto et al., 2014), but there is still scarce proof of their actual involvement and there is no direct evidence of the epigenetic changes that consequently occur in the developing fetus.

Clinical Evidence

Evidence of Across Indirect Epigenetics and Transgenerational Epigenetic Inheritance

Across indirect epigenetic changes, per se, define only intergenerational epigenetic inheritance, which is inheritance from one generation to the next ( Pang et al., 2021). AIE can be and has been considered, instead, a necessary but insufficient condition for transgenerational epigenetic inheritance, at least per its canonical definition ( Skinner, 2008). Many experiments have been performed to prove some form of epigenetic inheritance in the past two decades. We report several examples below.

Animal Models

Rat malnutrition impairs cognition in offspring ( Galler and Seelig, 1981). A low-protein diet over 10 generations produces even more severe cognitive deficits, which are evident after two generations, on returning to a regular diet ( Stewart et al., 1980). Dunn and Bale (2009) have demonstrated that a maternal high-fat diet in mice increases body size and insulin sensitivity, which endure until the second generation; these effects nearly vanish in the F3 generation, despite the alterations in body size being observed solely in female offspring, suggesting an imprinting mechanism. Parental addiction in rodents alter the sensitivity of offspring to drugs, eliciting adaptive counterregulatory responses ( Byrnes et al., 2011; Vassoler et al., 2013; Finegersh and Homanics, 2014).

Clinical Evidence

Direct proof of transgenerational epigenetic inheritance in humans remains lacking ( van Otterdijk and Michels, 2021). Nevertheless, there is notable indirect evidence (i.e., longitudinal studies with no or few insights into putative epigenetic mechanisms).


There is copious evidence of epigenetic changes in animal models, but this field must improve to generate stronger evidence and implement new techniques that could apply to human studies, in which direct and robust proof remains lacking. We have compiled many studies and pided them by epigenetic type and research model (Table 1). As discussed, this review’s aim is not to report all existing studies in the field but to provide some examples that can help us better understand epigenetic changes and their inheritance.

Table 1. Evidences of the three defined forms of epigenetic changes: direct epigenetics (DE), within indirect epigenetics (WIE) and across indirect epigenetics (AIE).

Epigenetic Mechanisms

How does epigenetic inheritance occur concretely? Although several epigenetic processes have been considered to answer this question, given the wide range of this work, we will focus on two of the more extensively studied mechanisms: methylation and ncRNA.

First-Generation Epigenetic Mechanisms

Methylation and Demethylation

DNA methylation is an enzymatic process by which a methyl group (CH 3) is covalently bound to the fifth position of a cytosine residue (5-methylcytosine, 5mC) to alter gene expssion. In mammalian DNA, this regulatory activity acts on CpG palindromes (i.e., diagonally symmetric couples of guanine-cytosine pairs), whereas asymmetric methylation is rare ( Chen and Li, 2004). When methylation affects the promoter region, it is associated with gene silencing-the most well-known function of this mechanism; however, when it involves the transcribed region, it increases transcriptional activity ( Jones, 2012). DNA methylation is involved in many processes, particularly those that are important for early development, such as genomic imprinting, X-chromosome inactivation, and transposon silencing ( Smith and Meissner, 2013).

The addition of CH 3 groups to CpG islands is catalyzed by DNA methyltransferases (DNMTs). DNMT1 primarily maintains DNA methylation patterns during replication, whereas DNMT3A, DNMT3B and DNMT3L (a noncatalytic isoform of DNMT3, termed DNMT3-like) are principally involved in establishing new DNA methylation patterns-a mechanism that is called de novo methylation-that characterize embryo development, in particular ( Chen and Li, 2004).

The maintenance of methylation is crucial for ensuring the continuity of the structural and functional identities of somatic cells throughout cell pision. During the S phase of the cell cycle, DNMT1 reaches hemimethylated CpGs with the aid of ubiquitin-like with PHD and RING finger domains 1 (UHRF1) proteins such that each newly synthesized DNA strand can be methylated per its complementary strand. Thus, after each replication, the symmetry of the methylation pattern is restored ( Zhang et al., 2011; Wu and Zhang, 2014).

Although methylation patterns are stable, they can be erased by two mechanisms: active and passive demethylation. Passive demethylation repsents a failure in maintenance (so-called replication-dependent dilution) and occurs primarily in the absence of functional DNMT1/UHRF1: if the symmetry of methylation is not reestablished, methylation is lost through replications ( Smith and Meissner, 2013; Wu and Zhang, 2014).

Figure 3. Methylation and demethylation. Methylation is a regulatory process of gene expssion, catalyzed by DNA methyltransferase enzymes, owing to the addition of a methyl group to the fifth position of a cytosine. DNA methyltransferases 1 (DNM1) is mainly involved in maintenance methylation that restores symmetric DNA methylation patterns after DNA replication. DNM3A, DNMTB and DNMTL, instead, are involved in the catalytic process that produces de novo methylation by adding methyl groups to unmethylated DNA strands. Methylation processes can be reverted by two mechanisms: passive demethylation due to loss of methylation across consecutive DNA replications; active demethylation mediated by ten-eleven translocation (TET) proteins.

Methylation and Epigenetic Inheritance

Maintenance and de novo methylation and active and passive demethylation are crucial for embryonic development and epigenetic inheritance. Gametes are completely demethylated and are remethylated after fertilization to erase all epigenetic marks that an inpidual accumulates over his lifespan. However, this resetting process is impeded during early development, perhaps accounting for transgenerational transmission of these epigenetic footprints ( van Otterdijk and Michels, 2021).

Elimination and restoration of methylation markers occurs in two steps (Figure 4). Immediately after fertilization, global demethylation is observed that erases methylation marks of the parental gametes through two sex-dependent mechanisms. First, the DNA in paternal pronuclei undergoes rapid, active demethylation that is mediated by TET3 proteins, which spare only imprinting control regions (ICRs) and certain retrotransposons, such as intracisternal A particles. This process takes place at approximately the time of DNA replication and ends before the first cell pision is completed. Then, the maternal genome is progressively demethylated through passive demethylation across subsequent cleavage steps ( Seisenberger et al., 2013). Consequently, the totipotency of the zygote is established and maintained across the first several cell pisions.

Figure 4. Biomarker reset. The elimination and restoration of methylation markers happen in two steps. A first, active demethylation takes place in parental gametes, right after fertilization. This process is mostly active-and therefore faster: it is completed by the first cell pision-for paternally inherited genome, while maternal pronucleus is slowly demethylated by passive diffusion across replications. This first global erasure of methylation marks spares only imprinted loci and some retrotransposons, and it is deemed to establish cellular totipotency. After the implantation of the developing blastocyst, a first de novo methylation wave begins, driving the crucial process of cellular differentiation. At the beginning of gametogenesis, when primordial germ cells start to migrate, a second demethylation takes place: gametes’ chromatin is globally demethylated, also including imprinted loci. After sex-determination, gametogonia are remethylated by a second wave of de novo methylation, which is higher (90%) and faster (it is mostly complete before birth) for male gametes and slower (40%) and lower (it does not end until puberty) for female gametes. Imprinting patterns are usually reestablished during this phase. The established patterns can be altered by direct or indirect experiences, particularly during gestation and right after birth. These processes depend on the activity of several epigenetic enzymes, among which DNA methyltransferases (DNMTs) and TETs are prominent. The regulation of these processes by non-coding RNA (ncRNA), has also been established.

The maternal factor Stella has been suggested to protect the maternal genome and paternal ICRs and intracisternal A particles from active demethylation. These regions undergo H3K9 (a Stella binding site) demethylation. Moreover, inside of the oocyte and zygote, the DNMT1o isoform pdominates and is more concentrated in their cytoplasm. In contrast, DNMT1 is the chief isoform in somatic cell nuclei but is scarce in the zygote. These differences in nuclear and cytoplasmic concentrations of DNMT1 isoforms account for global passive demethylation and might explain the maintenance of maternal ICRs ( Cardoso and Leonhardt, 1999; Seisenberger et al., 2013). Nevertheless, recent studies suggest that active and passive processes govern the demethylation of the maternal and paternal genomes ( van Otterdijk and Michels, 2021). After the implantation of the developing blastocyst, the inner mass cells (IMCs) undergo a wave of de novo methylation, which drives their differentiation. This process is mediated by DNMT3 ( Chen and Li, 2004; Seisenberger et al., 2013).

It appears that epigenetic transmission might be possible when the second demethylation step is pvented, as in the case of genomic imprinting, which constitutes the strongest evidence for transgenerational epigenetic inheritance in mammals ( van Otterdijk and Michels, 2021). Correct repssion of transcription of certain genes is crucial for a good developmental outcome. A glitch during genomic imprinting, for example, can cause severe pathologies, such as Prader-Willi and Angelman syndromes, which are derived from the loss of nonimprinted paternal and maternal genes, respectively ( Cassidy et al., 2000).

New-Generation Epigenetic Mechanisms

Non Coding RNA and Epigenetic Regulation

ncRNAs that are less than 200 nucleotides are labeled “short” or “small,” whereas those that exceed this length are defined as “long” (lncRNAs). These two groups can be subpided, depending on their genomic origin and biogenic activity.

lncRNAs are pided into five subgroups:

* natural antisense transcript (NAT), a complementary sequence to a coding RNA at the same locus (cis-NAT) or a distal genomic locus (trans-NAT).

* long intergenic ncRNA (lincRNA), which is encoded from the introns of intergenic regions (macroRNA or vlincRNA).

* sense overlapping, which is transcribed from the same DNA strand as another transcript.

* sense intronic, originating from the introns of coding genes.

* processed transcript, an RNA transcript that is spliced or polyadenylated.

Whereas NATs primarily regulate the expssion of the sense partner transcript, the activities of the other four classes remain unknown, but they are likely to include transcriptional regulation, RNA stability, and the recruitment of protein complexes and other subcellular elements. lncRNAs are usually transcribed and processed similarly to coding mRNAs ( Peschansky and Wahlestedt, 2014).

Small RNAs are grouped into five clusters: PIWI-interacting RNAs (piRNAs), endogenous short interfering RNAs (endo-siRNAs), miRNAs (or miRs), transfer-derived RNAs (tDRs or tsRNAs) and small nucleolar RNAs (snoRNAs). PiRNAs are usually composed of 26-30 nucleotides and can silence the transcription of target RNAs, promoting the trimethylation of histone 3 lysine 9 (H3K9me3), a marker of inactive chromatin, by a histone methyltransferase ( Luteijn and Ketting, 2013).

Our understanding of the processes that generate mature small ncRNAs is patchy. Only the biogenesis of miRNAs has been determined. The formation of miRNAs begins in the nucleus with the transcription of a primary miRNA (pri-miRNA) by RNA polymerase II (RNA Pol II). Pri-miRNAs are attacked by the microprocessor complex, composed of RNase III (Drosha) and DGCR8 (Pasha). Drosha cleaves pri-miRNA into a shorter transcript, whereas Pasha stabilizes the interaction between Drosha and pri-mRNA. This catalytic event produces a stem-loop structure, the pcursor miRNA (p-miRNA). The p-miRNA is then exported to the cytoplasm by Ran-GTP, which energizes the transport system, and exportin-5 (EXP5), which interacts directly with the stem-loop structure. Here, the p-miRNA associates with Dicer (another RNase III), which cleaves it into two molecules of approximately 22 nucleotides: guide strand (or mature miRNA) and passenger strand (or miRNA*). These two species are then loaded into argonaute (Ago) proteins, which select the mature miRNA (while miRNA* is degraded) and deliver it to the RNA-induced silencing complex, through which it arrives at its targets, destabilizing mRNA and inhibiting transcription ( Blahna and Hata, 2012).

ncRNAs can be epigenetic targets and epigenetic effectors. Their genetic loci can be subject to epigenetic regulation, like protein-coding genes, becoming susceptible of environmental influences; further, they govern gene expssion ( Peschansky and Wahlestedt, 2014; Szyf, 2021). This dual nature of ncRNAs implicates them as “change amplifiers.” In this sense, ncRNAs are similar to transcription factors.

Small RNAs and Epigenetic Inheritance

As discussed, fetal programming alone does not account for epigenetic transmission, unless we include the effect of pvious environmental factors (i.e., AIE) in its definition. As pointed out by Bohacek and Mansuy (2015), germ cell reprogramming could be a key mechanism of transgenerational epigenetic inheritance. Notably, miRNAs control de novo DNA methylation by regulating transcriptional repssors ( Sinkkonen et al., 2008). Epigenetic changes in germ cells arise and are maintained throughout methylation and acetylation, but miRNAs, particularly those in sperm, appear to have important functions (e.g., Bohacek and Mansuy, 2021; Rodgers et al., 2021; Fraser and Lin, 2021; Pang et al., 2021; Yeshurun and Hannan, 2021).

Conversely, global suppssion of miRNA (paired with the functional pdominance of endo-siRNAs) has been observed in mature oocytes and during early embryonic development ( Ma et al., 2010; Suh et al., 2010). Consistent with these data, oocytes lack DGCR8 (Pasha), which is necessary for miRNA but not endo-siRNA pathways ( Ma et al., 2010). miRNAs could be important mediators of placental development through their regulation of genetic expssion ( Babenko et al., 2021). Their function in the latter phases of zygote development remains unknown, but as we will discuss, there is evidence of the role of miRNAs in the regulation of oocyte function ( Tang et al., 2007; Soni et al., 2013). piRNAs are another class of small RNAs that are important in epigenetic inheritance and are highly expssed in sperm and oocytes; tsRNAs, which are enriched in mature mouse sperm, are critical in epigenetic inheritance ( Peng et al., 2012; Roovers et al., 2021; Chen Q. et al., 2021; Sharma et al., 2021). However, much work is needed to determine their functions.

Mechanisms of Epigenetic Inheritance: An Overview

NcRNAs might mediate the establishment of new patterns of gene expssion by regulating DNMT1 and TET in adult somatic cells (DE). Following fertilization, synchronous alterations to the intrauterine environment could define new expssion patterns (WIE), particularly through the activities of small ncRNAs on DNMT1, DNMT3A/B/L and TET, interfering with the maintenance of pexisting epigenetic hallmarks. Depending on when an environmental change occurs, the influence on the offspring might depend on the offspring’s sex and materialize using a sex-specific cluster of enzymes (see Figure 4).

Methodological Matters: Maternal vs. Paternal Contribution

The first studies on epigenetic forms of transmission focused on the effects of maternal care on the early stages of life, later considering nongenetic forms of developmental programming of fetal development during pgnancy. A practical problem arose, however: because mothers carry their children for 9 months and then care for them, it was difficult to distinguish between p-, peri- and postnatal epigenetic effects. Thus, several groups concluded that the paternal contribution should be considered. In many species, the only contribution of males is their sperm, which does not interfere with the gestational and postnatal periods.

This approach has been useful in demonstrating epigenetic inheritance, but it does not allow one to frame the entire landscape of mechanisms of epigenetic transmission: excluding maternal pgestational function because it is intractable for study fails to demonstrate that it does not exist or that it is irrelevant. Most of the literature has focused on the paternal role in mediating AIE (see Yeshurun and Hannan, 2021), whereas maternal function has been neglected. The drawback of many models of epigenetic inheritance is that they do not allow one to distinguish and define paternal and maternal contributions simultaneously for every effector that mediates the transmission of a certain property, such as stress reactivity. Stress vulnerability could result from the co-occurrence of maternal and paternal factors or show maternal or paternal pference, depending on the effector (e.g., which miRNA or group of miRNAs). Further, the pvalence of maternal and paternal contributions could depend on environmental conditions that could bring about, for example, paternal pvalence when the father is stressed or the pdominance of maternal contribution under baseline conditions.

Methodological Insights and Technical Niceties

Table 2. Some useful techniques that can be used to study and control for some crucial developmental variables.

Defining the Spacetime of Epigenetic Inheritance: Ideal Models

As reported above, several experiments have been conducted to demonstrate the existence of epigenetic inheritance. The results remain incomplete and sometimes conflicting, perhaps because only one route of transmission is usually considered at a time (e.g., maternal stress during pgnancy, paternal stress before mating). Moreover, the same type of event can occur in disparate moments and contexts, targeting subsequent generation through different routes.

This possibility implies that it would be better to apply several types of environmental conditions on all possible levels. For example, male and female mice could be stressed immediately prior to or long before fertilization-mildly or robustly and acutely or chronically-but also during gestation or after delivery (the latter two with regard to mothers only). It would then be interesting to study how a certain transmitted vulnerability interacts with an environmental condition that is similar to the causative factor throughout the offspring’s life. This approach is consistent with the model that, as in genetic inheritance, epigenetic inheritance can mediate the transmission of vulnerability (considered a type of epigenetic diathesis), which could remain silent and unexpssed unless-or until, depending on one’s degree of fatalism-certain environmental events take place ( Godfrey et al., 2007). Once epigenetic inheritance has been detected, the next crucial step is to determine the underlying molecular mechanisms.

The specific spacetime of an action of an epigenetic effector that is suspected to mediate transgenerational epigenetic transmission (for example, a miRNA) should be identified using the following experimental design. In a murine model, WT and manipulated (M)-i.e., KO, OE, or KD of the gene that encodes the epigenetic effector-oocytes could be fertilized with WT or M sperm in all possible combinations through IVF or natural breeding that is paired with embryo explants and implantation. The four possible types of zygotes that are produced could be implanted in WT or M dams-the latter of which allows one to control the effects of the intrauterine environment (including the placenta).

Figure 5. From in vitro fertilization (IVF) to fostering. Here, we schematize the suggested ideal model that could help define with great pcision the spacetime of a given epigenetic factor’s action. Once its role in fetal programming has been established, investigating its possible play in transgenerational epigenetic inheritance processes might be easier. See the text for more details.

Conditional models are pferred when defining the weight of a specific effector in a specific place and time (e.g., during paternal or maternal gametogenesis, zygote formation, the third week of gestation throughout the placenta, right after birth). In contrast, developmental models should allow one to observe the final, complex outcome of a certain alteration of a gene (such as polymorphisms and genes that encode epigenetic elements) in a more complex, systemic manner. The latter approach is not conducive to gaining a pcise understanding of mechanisms but still has ecological value that cannot be ignored.

Another noteworthy issue concerns whether to use IVF or natural breeding, followed by embryo extraction and implantation. IVF requires superovulation and the use of an artificial culture, which could alter the programing of gametes ( Bohacek and Mansuy, 2021). The use of natural breeding, conversely, fails to control for the effects of the manipulation of male and female reproductive fluids ( Bohacek and Mansuy, 2021), warranting further comparison with offspring that result from natural breeding.

These considerations are pivotal to correctly interpt data, despite the manipulation of a factor and the breeding procedure (e.g., conditional vs. developmental and artificial vs. natural). Moreover, the proposed model is only theoretical and does not impose its complete application, although it would likely produce the strongest evidence possible, whatever results emerge. Once the activity of a certain effector has been described, a more specific molecular analysis can be conducted to link the steps of the underlying mechanism of the specific process of epigenetic inheritance.

Conclusions and Future Perspectives

In this review article, we have introduced the concept of epigenetics, defining its spatial and temporal properties, allowing us to distinguish between types of epigenetics: a direct form of epigenetics (DE) and two forms of indirect epigenetics-within (WIE) and across (AIE). We have organized the main body of epigenetic evidence according to these three categories and focused on the latter (AIE), referring to it as a more rapid means of transmitting information across generations-compared with genetic inheritance-that guides human evolution in a Lamarckian (i.e., experience-dependent) manner. We have thus defined epigenetic inheritance in terms of AIE and illustrated the putative molecular mechanisms of this phenomenon.

Finally, we have discussed the main methodological matters regarding the study of epigenetic inheritance and have suggested strategies to solve some of the most compelling technical and theoretical problems that plague this field. The experimental models that we have proposed are inapplicable to human research, for obvious ethical reasons, but if we detail the mechanisms that underlie epigenetic inheritance, thus isolating key effectors to examine, we could study the “natural experiments” that we have (and probably will) occasionally encountered in history. There is no doubt that translational research could benefit from this scientific effort. Epigenetic inheritance, when maladaptive, can have a silent, unseen, but dramatic impact on health, perpetrating detrimental adaptations across generations.

Thus, there is no reason why a similar therapeutic approach should be overlooked for epigenetic abnormalities that affect an inpidual at early age and even during fetal development.

The environment is another level that confirms its well-established function as an epigenetic regulator and is also thus a potentially invaluable therapeutic “tool” ( Maccari et al., 2021). To strengthen the therapeutic power of the environment, paradoxically, we must understand the specific mechanisms that are altered by epigenetic adaptations following certain experiences. Yeshurun and Hannan (2018) have suggested a therapeutic/pventive approach, called “enviromimetics,” that aims to ameliorate paternal psychophysical conditions before conception to revert or pvent epigenetic alterations in sperm, thus reducing the transgenerational impact of stress.

Attaining this ideal therapeutic power will require new studies on AIE-particularly on the gap between two generations. These studies could ensure greater “spacetime resolution” of such a complex phenomenon, thus facilitating the development of a prompt and effective intervention. Although we have detailed how epigenetic factors can lead to many pathologies, we must be reminded that they are usually crucial in all of the adaptive processes that ensure the survival of the inpidual and species ( van Otterdijk and Michels, 2021). For this reason, the decision to interfere with their activity should be strongly supported by a profound understanding of the specific case in question and applied with great caution.

Author Contributions

IL conceived the general theoretical framework, collected most of the bibliography, wrote the first draft of the article and designed and drew the ps. Both authors developed, refined and carefully reviewed the final version of the article.


This work was supported by the Research Projects of Sapienza University of Rome grants ATENEO AA 2021 (C351BDB6) and ATENEO AA 2021 (RG11715C7E0A7187).

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.



Conradt, E., Lester, B. M., Appleton, A. A., Armstrong, D. A., and Marsit, C. J. (2013). The roles of DNA methylation of NR3C1 and 11β-HSD2 and exposure to maternal mood disorder in utero on newborn neurobehavior. Epigenetics 8, 1321-1329. doi: 10.4161/epi.26634

Cooper, G. M., and Hausman, R. E. (2013). “RNA synthesis and processing,” in The Cell: A Molecular Approach, eds M. Cooper and R. E. Hausman (Sunderland: Sinauer Associates, Inc.), 239-296.

Google Scholar

Crick, F. H. (1958). “The biological replication of macromolecules,” in Symposia of the Society Experimental Biology XII, eds F. Hildebrandt and P. Igarashi 138-163.

Dawkins, R. (1990). The Selfish Gene. New York, NY: Oxford University Press.

Johnson, T. D. (2010). “Developmental system theory,” in Oxford Handbook of Developmental Behavioral Neuroscience, ed. M. S. Blumberg, J. H. Freeman and S. R. Robinson (New York, NY: Oxford University Press), 12-29.

Radford, E. J., Ito, M., Shi, H., Corish, J. A., Yamazawa, K., Isganaitis, E., et al. (2014). In utero effects: in utero undernourishment perturbs the adult sperm methylome and intergenerational metabolism. Science 345:1255903. doi: 10.1126/science.1255903

【#5】Lộ Trình Trở Thành Frontend Developer

Tại sao tôi lại làm ra cái roadmap này ư? Tôi đã là một Fullstack Developer được 5 năm và hiện là kĩ sư lead cho tajawal . Nó không chỉ là sở thích mà cũng là một phần trọng trách của tôi phải nắm bắt các xu hướng và hướng dẫn các dev thật khác vững vàng. Tôi có thể thấy rất nhiều bạn mới vào nghề bối rối về việc bắt kịp tiến độ. Cũng rất nhiều bạn đã email hỏi xin lời khuyên về việc họ nên học cái gì từ năm 2021 đến nay. Vì thế cũng để tôi không bị “trùm bao bố” và giúp các bạn thì tôi đã quyết định làm các chart này 1 lần và mãi mãi.

Thật ra ban đầu chúng chỉ là một số đề xuất nhỏ về tool nhưng sau đó tôi muốn nó theo cấu trúc một tí, thêm một số chi tiết và thiết kế lại theo bước để bạn dễ hiểu hơn mà chọn và sắp xếp.

Trước khi đi giải thích từng điểm, tôi phải nói vài dòng trước:

Học một số cái căn bản của HTML

  • Học những cái cơ bản và học viết semantic HTML
  • Hiểu cách chia page thành các section và cách sắp xesp DOM hợp lý

Task -  Sau khi đã nắm những điều cơ bản về HTML, hãy tạo ít nhất 5 page HTML. Tôi khuyên bạn chọn bất kì website nào ví dụ, một profile page github bất kì hoặc login page của twitter. Sau đó tập trung vào nội dung. Có thể nó hơi xấu ban đầu nhưng đưunfg lo lắng quá và cứ tập trung vào structure chính.


Sau khi đã tạo được phần xương sườn của page, chúng ta sẽ cho nó một vài lớp áo và tô điểm cho nó đẹp một chút. Dùng CSS - Cascading stylesheets để làm đẹp cho các page HTML.

  • Đầu tiên phải biết về syntax và phải quen với các CSS property.
  • Học về box model và học cách chuẩn bị các layout bằng Grid và Flexbox
  • Sau khi đã xong, học cách làm responsive web bằng các media query.

Task — Một khi đã xong phần cơ bản, tiếp theo bạn cần phải chọn style cho các trang HTML bạn làm ở bước cuối. Ví dụ, nếu bạn viết trang HTML cho Github profile, bạn nên áp dụng CSS và làm cho nó trông giống như một Github profile page. Cứ làm thế với cả 5 page đã viết trước đó.

Học cái căn bản của JavaScript

JavaScript sẽ giúp bạn làm cho các trang HTML tương tác hơn. Ví dụ, với mọi slider, popup và noti bạn thấy trên website và khi nó reload một phần nhỏ của trang mà không phải reload nguyên trang, đó là do nó đã được xử lý bởi JavaScript. Trong bước này, bạn sẽ phải học JavaScript căn bản để chuẩn bị thật tốt

  • Học syntax và các cấu trúc căn bản.
  • Học cách thông thạo DOM bằng JavaScript ví dụ cách remove một số element khỏi page, cách add thêm element, add hoặc remove các class, apply các style của CSS,… bằng JavaScript.
  • Học cách tạo các HTTP call bằng XHR hoặc Ajax. Ajax sẽ giúp bạn thao tác mà không phải reload page.
  • Tiếp đến hãy học về feature mới ES6+. ES6 là một phiên bản JavaScript gồm nhiều update mới như classes, nhiều cách để declare các variable, add thêm các method mới vào array, string concatenation v.v. Hầu hết các bài viết bạn đọc online họ sẽ dùng để giải thích về ES6 – một transpiler mà convert JavaScript thành vì nó không được các browser cũ support. Nhưng khoan hẵn lo về Babel, chỉ cần hiểu trước ý tưởng và dùng nó cho cac browser mới hơn có support ES6. CHúng ta sẽ quay lại ES6 sau.

Có nên học jQuery không?

Có một thời mà ai cũng phát cuồng về jQuery, và thực sự thì nó là một library rất quyền lực cung cấp vỏ bọc cho JavaScript và để bạn thao tác mọi thứ trên browser một cách thuận tiện nhất. Nhưng ngày đó đã qua đi, bây giờ người ta đã không còn dùng nó quá nhiều trong các project mới nữa nhưng đương nhiên vẫn còn người dùng nó. Không nhất thiết phải học nhưng nó khá dễ và cũng có ích nếu bạn có tìm hiểu.

Thực hành

Đừng học nếu không thực hành, chứ không thực hành thì tôi không hiểu bạn nhớ được cái gì nữa. Lúc mới học xong thì có thể bạn sẽ nhớ và hiểu sâu hiểu xa, nhưng đó chỉ là nhất thời và sẽ bay đi đâu mất ngay. Hãy thực hành càng nhiều càng tốt trong suốt quá trình này.

Hãy tiếp tục là một số các responsive web và add thêm tính tương tác bằng JavaScript. Bạn cũng có thể copy một webpage bất kì mà bạn thấy hay ho nhưng hãy nhớ sử dụng hết những thứ mình đã học đến thời điểm hiện tại.

Sau khi đã tạo được vài website thì bây giờ sẽ bắt đầu công việc thực sự. Vào và search thêm project và mở một số pull request trên các project nguồn mở. Một số gợi ý cho pull request như sau:

    Đầu tư thêm vào UI, tạo thêm các trang demo hoặc cải thiện phần design
  • Xem các vấn đề mở mà bạn có thể giải quyết
  • Refactor code bất kì mà bạn nghĩ là cần cải thiện

this repository này , nói rằng bạn đang học và xin feedback cho phần PR cũng như bạn có thể cải thiện phần nào.

Ở phần này bạn cũng cần có một số kiến thức về git. Bạn có thể không làm nhưng nếu có thì tôi cam đoan là bạn sẽ không hối hận, bạn sẽ bị sốc khi thấy có bao nhiêu người sẵn lòng giúp đỡ. Bạn có thể tìm thấy rất nhiều resources git miễn phí, ví dụ như .

Tự tạo động lực cho bản thân

Nếu học mọi thứ đàng hoàng, bạn hoàn toàn có thể tìm một số job freelance hoặc thậm chí cả fulltime. Tuy nhiên, đừng dừng lại ở đó, chặng đường vẫn còn dài nếu như bạn muốn một công việc thật ổn định và mỹ mãn.

Package Managers

Sử dụng hết những cái đã học

Để chuẩn bị cho bước này, nếu trước đây bạn dùng các thư viện ngoài ví dụ một plugin hoặc một widget ngoài bất kì, bạn phải down các file JavaScript và CSS thủ công và để vào trong project. Các thư viện hoặc plugin nếu ra version mới cũng phải update ngay và chèn vào project, một khâu hết sức phức tạp. Thì package manager sẽ giúp bạn dỡ bớt sự phức tạp này khỏi workflow. Nó sẽ tự mang các library và plugin ngoài vào project của mình, và bạn sẽ không phải lo về việc copy các library thủ công nữa hay phải update phiên bản mới. Có sẵn yarn và npm . Cả hai khá giống nhau, nên bạn có thể chọn cái nào cũng được, học một cái thì cũng sẽ biết dùng cái còn lại.

CSS Preprocessors

Preprocessors đã trau dồi thêm cho CSS với rất nhiều tính năng mà không có mặc định. Có rất nhiều option như Sass, Less, Stylus v.v. Nếu tôi phải chọn một, thì Sass là câu trả lời. Tuy nhiên, PostCSS gần đây thu hút khá nhiều sự chú ý, nó như một dạng “Babel” của CSS. Bạn có thể chỉ dùng nó hoặc cùng Sass. Tôi khuyên bạn nên học Sass trước và tìm đến PostCSS sau nếu có thời gian.

Các framework CSS

Bạn không cần phải học framework nữa, tuy nhiên nếu vẫn muốn thì có rất nhiều lựa chọn cho bạn. Qua những cái đã thử tôi thích nhất là Bootstrap, Materialize và Bulma. Nhưng nếu nhìn vào thị trường hiện tại, tôi sẽ chọn Bootstrap để bắt đầu .

Sắp xếp CSS

Khi app càng ngày càng lớn, thì CSS cũng trở nên bừa bộn hơn và khó quản lý. Có rất nhiều cách để kiến thiết lại CSS, ví dụ, OOCSS, SMACSS, SUITCSS, Atomic và BEM. Bạn nên tìm hiểu sự khác nhau giữa chúng, nhưng cá nhân tôi thích hơn.

Build tools

Các tool giúp bạn build/bundle và làm app JavaScript. Loại này bao gồm các linter, task runner và các bundler.

Đối với các task runner, trước đây có rất nhiều lựa chọn như npm scripts, gulp, grunt v.v. Nhưng hiện tại thì webpack đã giúp bạn xử lý gần như mọi thứ mà bạn làm với gulp, Chỉ cần duy nhất một npm script trong task runner dùng để tự động hoá các task mà webpack có thể làm. Không nhất thiết phải học Gulp, tuy nhiên sau này nếu bạn có thời gian, cứ tìm hiểu thêm và xem xem liệu nó có giúp được gì cho app của bạn không.

Về các linter, cũng đã từng có rất nhiều option như ESLint, JSLint, JSHint và JSCS. Nhưng hiện tại ESLint làm gần như mọi thứ.

(tiếp tục) Thực hành - Tạo ra cái mới

Đối với module bundler, ngày xưa thì có Parcel, Webpack, Rollup, Browserify v.v. Bây giờ nếu phải chọn thì hãy chọn Webpack. Rollup cũng khá phổ biến nhưng nó nên dùng cho các thư viện; còn khi nhắc đến app, đã có webpack. Thế nên hãy tự học webpack và thăm lại section này bằng Rollup nếu muốn.

Chọn một Framework

Trong roadmap cũ thì section này thường ngay sau phần kiến thức căn bản nhưng tôi đã dời nó lại sau Sass, buils tool và các package managers vì bạn sẽ dùng hết chúng trong framework và nếu không học trước thì vô dụng cả thôi.

Lựa chọn thì vô vàn. Dạo này thì có React, VueAngular là phổ biến nhất . Nhu cầu về chúng tôi thì càng ngày càng nhiều. Bạn có thể chọn bất kì cái nào bạn muốn, cá nhân tôi thiên về React hoặc Angular. Có một lưu ý nhỏ ở đây, nếu là newbie, có thể bạn sẽ thấy Angular tương đối đơn giản, căn bản là vì nó support gần như mọi thứ ví dụ powerful router có support tình trạng lazy loading, HTTP client có support các interceptor, dependency injection, component CSS encapsulation v.v mà không phải lo về các thư viện ngoài. Tuy nhiên React chắc chắn cao tay hơn trong việc tìm kiếm cộng đồng quanh nó và đội ngũ tại Facebook đang làm việc để cải thiện nó. Hãy đảm bảo rằng bạn không chọn nó vì độ hot, mà chọn cái nào phù hợp nhất với mình.

Nếu bạn chọn chọn Angular thì bạn cần thực hiện TypeScript (thực ra bạn không cần đến nó khi là app Angular nhưng có nó thì tốt hơn) và Rx.js cũng sẽ giúp ích cho app của bạn nhiều. Đây là một thư viện khá mạnh và phù hợp với functional programming.

Và nếu bạn chọn Vue.js, bạn sẽ cần học Vuex tương tự như Redux dùng riêng cho Vue.

Lưu ý một điều là Redux, Mobx và chúng tôi không ràng buộc với các framework này, bạn hoàn toàn có thể dùng nó cho app JavaScript nữa. Đồng thời, nếu bạn chọn Angular, nhớ là dùng Angular 2+ chứ không phải Angular 1+

Thực hành-time

Giờ đây bạn đã biết *mọi thứ* cần để build được một app JavaScript. Hãy tạo thêm app từ framework bạn đã chọn. Bạn có thể lấy ý tưởng từ ideas directory trên repository rồi bắt đầu ngay.

Progressive WebApps

Khi đã done bạn có thể tìm hiểu về các service worker cũng như cách tạo nên progress web app.


Tuỳ vào mục đích mà có rất nhiều tool để bạn chọn. Tôi thuường dùng tổ hợp Jest, Mocha, Karma và Enzyme. Tuy nhiên, phải tìm hiểu và biết điểm khác nhau giữa chúng là gì để chọn được cái thích hợp nhất.

Static Type Checker

Static type checker giúp bạn add thêm type checking vào JavaScript. Bạn không cần đến chúng nhưng chúng thật sự rất hữu ích và chỉ cần vài tiếng tập trung là bạn có thể hiểu nó. CHủ yếu nó là TypeScript Flow. Tôi rất thích TypeScript và sẽ chọn nó trước, nhưng bạn có thể check hết cả 2 để xem thử.

Server Side Render

Đừng chỉ dừng ở Checker vì đây là một phần cần thiết trong lộ trình làm Frontend của bạn.

Hãy tìm hiểu thêm về server side rendering trên framework bạn đã chọn. Ví dụ, nếu bạn chọn React, thì có và rất phổ biến. Đối với angular thì có . Và với chúng tôi đã có .

Chắc chắn vẫn còn một số điểm chưa được nhắc đến trên roadmap nhưng căn bản đây là những gì bạn cần để trở thành một “Frontend Engineer”. Và đừng quên, điểm quan trọng nhất đó là, thực hành nhiều nhất có thể. Ban đầu sẽ còn bỡ ngỡi và nhiều vấn đề, nhưng theo thời gian và luyện tậo bạn sẽ quen dần và lên tay rất nhanh thôi.

Bài viết gốc được đăng tải tại Medium

Happy coding!

Ban Biên Tập Blog TopDev. Nice to meet you