
Evo 2 is a genomic foundation model from Arc Institute for prediction and design across DNA, RNA and proteins. It models sequences at single-nucleotide resolution.
Web appFree
Step 1: Read the Evo 2 overview on the Arc Institute page.
Step 2: Open the Evo Designer to try sequence design.
Step 3: Follow the get started guide for Evo Designer.
Step 4: Get the model code from the Evo 2 GitHub repository for your own work.
Step 5: Use the interpretability visualizer to explore what the model learned.
It performs generalist prediction and design tasks across DNA, RNA and proteins.
It has 40 billion parameters and a 1 megabase context, trained on over 9 trillion nucleotides.
The page links to Evo 2 on GitHub and Evo 1 on Hugging Face.
Arc Institute. The page says Evo is a trademark of Arc Research Institute.
EvoBiological foundation model for sequence generation and prediction across DNA, RNA, and protein-related genomic tasks
Universal Cell EmbeddingsUniversal Cell Embeddings is a PyTorch implementation of the UCE model for embedding single-cell RNA sequencing datasets. It adds a cell embedding to each cell in an AnnData file.
CellPLMPretrained cell language model for single-cell embeddings, annotation, perturbation, and spatial transcriptomics tasks