"Accelerating Scientific Research" — Claude Across Disciplines
Anthropic has published a feature post documenting how researchers across disciplines are using Claude to accelerate discovery: genomics teams are cross-referencing whole-genome studies with published literature using the 1M-token context window; synthetic-biology labs are using Claude to design and verify experimental protocols; climate modellers are querying datasets through MCP connectors. The post also serves as a soft announcement that Claude now has direct connectors to PubMed, clinical trial databases, and several environmental data APIs.
One highlighted enterprise example is Novo Nordisk's "NovoScribe" AI platform, which uses Claude to draft regulatory documentation — condensing a weeks-long manual process into a reviewed draft in hours. NovoScribe was developed entirely in-house using the Claude API and MCP server tooling, with no custom model training required.
Best practice for scientific use cases: When using Claude to analyse research literature, pass the full text of the papers rather than abstracts — Opus 4.6's 1M context window makes this feasible, and Claude's synthesis quality improves significantly with access to methodology sections and supplementary data that abstracts omit.