SoftBank has led a $130m investment in a US bioengineering startup that makes “designer microbes”, in the latest sign of artificial intelligence and robotics invading the natural sciences.

软银(SoftBank)牵头向美国一家生产“设计师微生物”的生物工程创业型公司投资1.3亿美元，这是人工智能和机器人进军自然科学的最新迹象。

The three-year-old venture, Zymergen, uses machine learning and other techniques to re-engineer the genetic make-up of micro-organisms. The enhanced microbes are used in existing industrial processes including making generic pharmaceuticals, but its backers hope the technology will also open the way to bigger breakthroughs.

创立三年的Zymergen利用机器学习等技术重新设计微生物的基因构成。得到强化的微生物正用于包括生产仿制药在内的现有工业流程，但Zymergen的投资者希望，这项技术还能开辟出取得更大突破的道路。

Deep Nishar, head of the SoftBank’s new investments group, said the gene-editing of microbes could eventually be used to create new materials, such as adhesives that work in extreme conditions or flexible electronics for consumer gadgets that wouldn’t break when dropped.

软银新成立的投资集团的负责人迪普•尼沙尔(Deep Nishar)表示，微生物基因编辑最终可用于创造新材料，比如工作于极端条件下的粘合剂或柔性电子器件，后者可用来生产掉落后不会破碎的消费类电子产品。

Zymergen is part of a new generation of companies trying to use advanced computing to enhance bioengineering.

Zymergen是新一代试图利用先进计算来增强生物工程的企业之一。

Steven Chu, a former US energy secretary and Nobel prizewinner who is joining the Zymergen board, said that re-engineering the genomes of microbes had so far proved more difficult than anticipated, despite their relatively simple composition.

前美国能源部长、诺贝尔奖得主朱棣文(Steven Chu)也将加入Zymergen董事会。他表示，迄今为止的事实证明，重新设计微生物基因组比预期的要难，尽管它们的构成相对简单。

“You have to find a better way to programme the gene set,” Mr Chu said. He described the type of computing used by Zymergen as “the beginning of a new type of chemistry”.

朱棣文说：“你必须找到更好的方法来编辑基因组。”他说，Zymergen采用的那种计算“开启了一种新型的化学”。

The company uses machine learning — a form of advanced pattern recognition — to try to identify the groups of genes inside microbes that are likely to cause a desired result, such as producing a certain protein. That is designed to replace current approaches largely based on trial and error and could yield far more efficient organisms.

Zymergen使用机器学习（一种先进的模式识别）来试图找出微生物内部可能导致所需结果的基因组，比如能产生某种蛋白质的基因组。该方法旨在取代目前主要基于试错的方法，有可能生成有效得多的微生物。