Combining artificial intelligence (AI) and gene editing allows us to modify bacteria and design proteins that can then be used to reduce greenhouse gasses, digest plastic, and replace agricultural pesticides. Combining the two, we can improve human health, conquer diseases once thought incurable, and develop lifelong personalized medical treatments. We can even develop lightning-fast responses to incipient pandemics.
AlphaFold came out in 2018. This machine-learning AI has been folding proteins since it first was developed. It has beaten human competitors in protein-folding contests. Since it first came on the scene it has gone through several iterations with AlphaFold 3, the latest to come out this year. In this new release, it can predict DNA and RNA protein complexes.
AlphaFold was used to model the COVID spike protein helping vaccine researchers to come up with mRNA vaccines to immunize the world. Currently, it is being used to design drugs that interact with target proteins and produce antibody therapies.
AlphaFold is being used to study proteins responsible for antibiotic resistance to fight superbugs. It is also modelling proteins associated with Parkinson’s and other neurological disorders.
AlphaFold researchers are engineering enzymes making biofuels, breaking down waste, and developing biomaterials for medical and industrial use.
Now, AI and CRISPR, the revolutionary genome editing tool are combining forces to produce proteins and bacteria that can breathe in carbon dioxide (CO2) and methane (CH4), the two most potent greenhouse gasses, and exhale organic feedstocks, or produce plastics as a substitute for concrete.
How hot is this subject area? David Baker at the University of Washington’s Institute for Protein Design is this year’s recipient of the Nobel Prize for Chemistry because of his work in this field. The Institute has been designing and predicting protein structures that are the building blocks of life. Baker isn’t alone in marrying CRISPR and AI. Evolutionary Scale, a biotech startup has trained a large language model to help design proteins. Called ESM3, it is the first generative model to create novel proteins. Another biotech company, Profluent, has developed OpenCRISPR-1, an AI-designed gene editor that reduces off-target effects by 95% compared to traditional CRISPR methods. Off-target effects are unintended DNA modifications that can occur when using CRISPR to edit DNA and RNA. Then there is ProGen, a Salesforce AI Research spinoff that in 2020 developed an advanced language protein-building model. The current release, ProGen2, has been trained on a dataset of over 280 million protein sequences.
What will the marriage of AI and CRISPR produce? Expect to see an acceleration in the pace of genetic research as AI is used to analyze large volumes of genomic data. AI will help to identify gene functions and predict gene mutation effects. AI will generate new versions of CRISPR not found in nature creating gene editors with capabilities that go far beyond our current technology.
Where will we see the benefits of combining AI and CRISPR? For medicine, AI-enhanced CRISPR will improve diagnosis and even prevent genetic disorders. It can make medicine even more personalized by analyzing an individual’s genetic profile and suggesting treatment. It can even train the immune system to recognize and attack different cancer cells.
Outside of medicine and materials science, combining AI and CRISPR will have a big impact on agriculture. It will help improve crops to make them more flavourful and nutritious while also improving disease, insect and drought resistance. It will have a similar impact on livestock.
Ethical concerns remain in combining these two powerful technologies. There have to be guardrails to stop mad scientists from creating genetic inequalities that lead to the development of designer babies and superhumans. Remember Star Trek’s, the Wrath of Khan, and you get the picture.