SynBio4: Marriage of SynBio and AI (v1.0)
Synthetic biology and artificial intelligence revolutions are proceeding in parallel and are overlapping in major ways.
In the mid 90's a field called bioinformatics came into existence to manage the growing databases of information on genes, proteins and other data intensive aspects of biology. It is an interdisciplinary field of science that develops methods and software tools for understanding biological data, especially when the data sets are large and complex. There was an increased use of mathematics and computation to understand this data. Computational approaches to biology are now being integrated with AI. More lately, AI is being used to process and make sense of the complexity of our biological experiments, make suggestions to fine tune our biological designs, controlling biological robots, and attempt to predict the outcome of biological experiments. These capabilities will only be enhanced in the future with more powerful computers and better algorithms and AI capabilities.
AI Deep learning has revolutionized protein structure modeling including protein folding in three dimensions. Proteins play an extremely huge role in our bodies and its function is highly dependent on its structure. Remember also the genetic code in DNA is an ordered list of four proteins and RNA are transcribed from DNA. A transcribed RNA called an mRNA transcript escape the nucleus and travel to the ribosomes, where they deliver their protein assembly instructions. AlphaFold and RoseTTAFold advance way beyond classical physically based approaches to protein structure prediction, and many areas of structural biology are likely to be affected by further advances in deep learning. AlphaFold is an AI program developed by DeepMind, a subsidiary of Alphabet (parent of Google), which performs predictions of protein structure. The program is designed as a deep learning system. Over 200 million proteins have been mapped by AlphaFold. RoseTTAFold is a software tool that uses deep learning to quickly and accurately predict protein structures based on limited information. It can be used to compute hundreds of new protein structures, including many poorly understood ones. Without the aid of such software, it can take years of laboratory work to determine the structure of just one protein. However, with RoseTTAFold for example, a protein structure can be computed in as little as ten minutes on a single gaming computer.
Currently, there are a number of open-source AI resources that scientists can use to design, edit and test gene sequences in silico. Scientists can and should use these tools to accelerate their research. But the freedom to experiment with and test new theories, ideas, and solutions comes with risks. It will be important to maintain a healthy awareness of the potential hazards, as well as what safeguards might look like in a space where AI is routinely deployed. It will therefore be key to proactively build in biosafety and biosecurity requirements into the review and design of new life sciences innovations. There is a lot of focus by policy makers on AI, especially in EU. But given that fusing AI with synthetic biology is still an emerging field, there is currently less attention from policy makers in this area.
How will the union of synthetic biology and AI factor into our lives more broadly? AI is the perfect tool for understanding the complexity of biology including deciphering protein structure (Knowledge of existing protein structures will allow us to arrive at new proteins that are useful), constructing or predicting gene sequences for potential immunotherapies, and contriving molecular formulas for synthesis of biomedicines. There is no telling what will emerge in this rapidly growing area in the future!!
On a lighter note, here is a take on AI by Jon Stewart!!
Comments