Protein-Design Tools

Download CSVs ▾

Built client-side from the JSON embedded in this page — no server, no network.

    built with openpulse.science

    Release timeline

    Spotlight: BindCraft

    tool · contributors · main author

    A deeper look at one tool: BindCraft (one-shot de-novo protein-binder design), its repository signals & contributors, its development activity over time (CHAOSS commit history), and the publication record of its main author — Martin Pačesa — enriched live from OpenAlex.

    Field at a glance

    The shape of the corpus — how the — tools split across research categories, implementation languages, open-source licenses, and maintenance status. Each ring totals the tools it could classify; hover a slice for the tools in it.

    Community health (CHAOSS)

    CHAOSS metrics reference ↗

    Live community-health metrics from the Open Pulse CHAOSS Metrics API — computed over Neo4j + Oxigraph + OpenSearch for the protein-design-tools project (74 seed repos). The focus set is grouped into the three CHAOSS topics: Community (alive & kicking?), Popularity (who uses & reuses it?), and Quality (can others understand & reuse it?). A value of — means no data (never zero). The commit-derived explorer charts further down come from the GrimoireLab git enrichment.

    Commit-derived signals · GrimoireLab / OpenSearch explorer

    Hover a tool to trace it across every chart · click to pin · double-click to open.

    Relationship explorer

    Each dot is a tool, colored by category. Dashed lines mark the median of each axis (four quadrants). Count axes use a log scale automatically. Hover/click a dot — it links to the bar charts below.

    Metrics over time

    One line per tool, from per-commit timestamps in OpenSearch (GrimoireLab). Hover or pin a tool to trace its trajectory — pins are shared with every chart on this page.

    one bar per tool · colored by category

    Pick a metric from the dropdown — the bar chart compares every tool, the panel beside it explains the metric and its leaders. Tools with no GrimoireLab commit data are omitted from commit-derived metrics only.

    Community engagement

    social edges · Open Pulse / Neo4j

    How people interact with each repository beyond starring it — opened PRs, opened issues, PR reviews and comments, counted as distinct interaction edges captured during the crawl. Hover a bar for the distinct-user count behind it; click to open the tool.

    Related publications

    enriched via OpenAlex ↗

    Each tool's canonical / landmark paper plus any schema:citation links in the graph, deduped by DOI and enriched live from OpenAlex (citation count, venue, open access, topics). Papers marked curated were found by web research (the graph has no citation for that tool) — verify the DOI before trusting it.

    Tool lineage

    A family tree of the field: each tool sits at its release year, sized by citation impact (OpenAlex) and colored by category. Curved links are "builds on" relationships — curated from the literature (e.g. AlphaFold 2 → ColabFold, RoseTTAFold → RFdiffusion, ProteinMPNN → LigandMPNN). Hover a node for detail.

    Downstream adoption

    Repositories that depend on each tool — incoming DEPENDS_ON edges from the crawl (which captured dependents). Click a bar to list a sample of the dependent repos.

    Top contributors

    People ranked across the 22 protein-design tools. Open Pulse footprint = total repos a person has contributed to platform-wide (op:hasContribution); followers / repos owned come from the enriched GitHub profile. Hover a name for the full card.

    🤗 Hugging Face for Science

    ⚑ added by AI agent · pending validation

    Hugging Science is a new community hub aggregating open AI models & datasets for research. Tools in this study publish models, interactive Spaces and datasets there.

    Related datasets & models in the field

    Not tied to a single tool — benchmarks and general protein language models the corpus is evaluated/built on.

    Contributor-overlap network

    Edges connect tools that share any author in their op:Contribution records. Edge thickness is the number of shared contributors. Every repository is shown — tools that share no contributor with any other appear as faded, dashed, unconnected nodes. Force: drag nodes to pin them, click to open the modal. Hive: each axis is one of the 5 categories ("waves" from the Step-1 brief), nodes ordered by stars along the axis.

    View

    Cross-tool contributors

    Authors who appear in two or more tools. ORCID identities are mapped to GitHub via schema:url. Affiliations come from org:hasMembership (ROR only). Each tool chip is colored by category — click a chip to open that tool's modal.

    Cross-tool organizations

    ROR-identified institutions reached by following each contributor's org:hasMembership → org:organization chain (with the ORCID → GitHub bridge applied). Only institutions shared by two or more contributors are listed — a single affiliate isn't a cross-tool signal. The Type column is color-coded to match the collaboration map.

    Institutional collaboration map

    Which organizations stand behind these tools, and how. Nodes are tools, ROR institutions and GitHub organizations / accounts; shape encodes the entity kind, color the organization type, and the line style the relationship — solid = owns the repo, dashed = a contributor is affiliated there (thicker = more shared contributors). Switch to Collaboration to see institutions linked when they share a contributor — that view clusters the field by who works with whom. Drag to pin, hover for detail, click a tool to open it.

    How this map is built · — snapshot
    Show

    Where it comes from

    The ROR institutions behind these tools, placed by their registered coordinates. Each marker is an institution — sized by how many tools it's tied to and colored by organization type (same palette as the collaboration map). GitHub-only owners and accounts aren't geolocated. Hover a marker for detail.

    Dependency ecosystem

    Outgoing DEPENDS_ON edges from the seed tools, per Neo4j. Seed nodes are colored by category; dependency-only nodes are gray and sized by how many seeds depend on them. Hover for the seeds sharing each library; click a node to open its modal (when it's also a seed).

    View