By Tom Chadwick, Account Executive, Copoly

The field of biomedical research is on the cusp of a transformative era, driven by the convergence of artificial intelligence (AI) and genomics. At the forefront of this revolution is EVO-2, a groundbreaking AI model developed by Arc Institute and NVIDIA. Trained on an unprecedented scale of over 9 trillion DNA base pairs from over 128,000 species across all domains of life, EVO-2 possesses a profound understanding of the genetic code. This deep learning model goes beyond simply analyzing existing genomes; it can generate entirely new genetic sequences, predict the impact of mutations, and even design artificial genomes.  

EVO-2’s Significance in Biomedical Research:

EVO-2’s capabilities have significant implications for various areas of biomedical research:

• Variant Impact Prediction: EVO-2 can accurately predict the functional consequences of genetic mutations, including those in non-coding regions previously considered “junk” DNA. This ability to identify disease-causing mutations with over 90% accuracy in genes like BRCA1, associated with breast cancer, can accelerate the development of targeted therapies and personalized medicine.  
• Genome Design: EVO-2 can generate complete genome sequences, including those for eukaryotic organisms like yeast and even small bacteria. This opens up new possibilities for synthetic biology, allowing researchers to design and engineer new biological systems with specific traits for research or therapeutic purposes.  
• Protein Engineering: By understanding the relationship between genetic sequences and protein structure and function, EVO-2 can aid in the design of novel proteins with desired properties. This has applications in drug discovery, where researchers can design proteins to target specific disease pathways or develop new enzymes for industrial biotechnology.  
• Accelerating Research: EVO-2 can significantly reduce the time and resources required for traditional experimental research. By predicting the effects of genetic variations, researchers can prioritize experiments and focus on the most promising candidates, leading to faster discoveries and advancements in the field.  

EVO-2’s Role in Accelerating In Silico and Preclinical Research

EVO-2 is not just an analytical tool; it is a powerful engine for accelerating both in silico and preclinical research. By integrating multi-omics data, EVO-2 can refine experimental designs, optimize target selection, and enhance the interpretation of complex genomic datasets.

• CRISPR Guide Design: EVO-2 can be used to generate optimized guide RNA sequences for CRISPR-based gene editing by predicting the most effective and least off-target sequences. This ensures higher precision in functional genomics screens and improves the efficiency of genome editing experiments.
• Mutation Analysis from NGS Data: Following next-generation sequencing (NGS) studies, EVO-2 can prioritize functionally significant mutations by predicting their impact on gene expression, protein structure, and disease relevance. This capability helps researchers focus on the most biologically relevant variants for further study.
• Multi-Omics Integration for Preclinical Models: EVO-2 can process vast multi-omics datasets—genomic, transcriptomic, epigenomic, and proteomic—to generate predictive models of disease progression and therapeutic response. By simulating gene interactions at a systems biology level, EVO-2 enables researchers to identify key biomarkers and therapeutic targets with greater confidence.

EVO-2’s Unique Architecture:

EVO-2’s remarkable capabilities are attributed to its unique architecture called StripedHyena 2, which allows it to process vast amounts of genetic data efficiently. This architecture enables EVO-2 to analyze sequences up to 1 million nucleotides long, providing a broader understanding of genomic interactions and enabling the study of complex biological processes.  

Accessibility and Future Directions

The developers of EVO-2 have made the model’s code, parameters, and training data publicly accessible, fostering collaboration and accelerating research in the field. This open-source approach encourages researchers worldwide to utilize and build upon EVO-2’s capabilities, leading to further advancements in AI-driven genomics and personalized medicine.

Bringing EVO-2 into Your Research Workflow

At Copoly, we specialize in integrating AI-driven tools like EVO-2 into biomedical research workflows. Whether you are looking to incorporate AI into your experimental design, streamline CRISPR-based gene editing, or analyze complex multi-omics datasets, our team can provide tailored solutions to enhance your research.

We work with academic institutions, biotech startups, and pharmaceutical companies to implement AI-powered approaches that optimize discovery pipelines, improve prediction accuracy, and accelerate translational research. If you’re interested in leveraging EVO-2 or other AI-driven tools for your research, contact us to explore how we can help you drive innovation and uncover new insights in genomics and precision medicine.