Boston Computational Biology Hackathon
Hosted by Anthropic, Modal, and Flagship Pioneering
Spend a day building computational biology tools and workflows with the Anthropic, Modal, and Flagship Pioneering teams.
The challenge:
Build a tool for running computational biology workflows in the cloud with Modal x Claude Science: autoresearch, molecular design, data scraping and processing, or any computationally heavy workload you want to automate with agents.
Inspiration:
Build beyond the bench: start from a user in the life sciences you can name – a scientist, a clinic, a biotech – and use Claude Science to create the tool they're missing.
Protein function: Run large protein language models, structure prediction, and large-scale homology/domain searches to characterize unknown proteins.Input: Protein sequence (FASTA).
Output: Predicted function, domains, homologs, structural insights, and confidence scores.
Drug discovery: Integrate large multimodal datasets and run large-scale target scoring, virtual screening, or model inference across many genes and compounds.Input: Disease, phenotype, or gene set.
Output: Ranked targets or drug candidates with scores and supporting evidence.
Imaging & phenotyping: Process high-content microscopy datasets with segmentation, image embeddings, phenotype discovery, and large-scale inference across millions of cells or images.Input: Microscopy images.
Output: Detected phenotypes, cell/image embeddings, segmentation results, and QC metrics.
Biology knowledge graph: Build and analyze large biological graphs using entity embeddings, multi-hop graph search, graph algorithms, link prediction, and LLM-based reasoning over large numbers of relationships.Input: Gene, protein, disease, drug, or biological question.
Output: Predicted relationships, ranked hypotheses, relevant graph paths, and supporting evidence.
What you'll get:
Modal and Claude credits to fuel your build
Direct access to the teams behind Modal, Claude Science, and Flagship Pi