From patents and papers to loadable SAR

Move from scattered scientific evidence to SAR your team can use

Excelra finds the evidence that matters, turns complex patents and papers into traceable records, and delivers data aligned to your program and informatics environment. Start with mining, curation, or an existing GOSTAR slice to reduce manual search and cleanup while giving scientists a stronger basis for modeling and decisions.

See which solution fits Describe your data need

Scientist-led screeningInclusion criteria guide every search

Three-tier scientific QCCuration, review, and senior QC

Patent-informed coverageUSPTO · EPO · WIPO

Schema-aligned deliverySDF · CSV · TSV · Excel · JSON · custom

Choose the right starting point

Clear the data bottleneck holding up the next decision

Confirm that sufficient evidence exists, convert known sources into usable records, or license a GOSTAR slice that is already curated. Each route removes a different source of delay, and they can be combined when the program requires it.

Coverage is uncertain

Data mining

Avoid funding a low-yield dataset build

Scientific search specialists work across patents, journals, and accessible grey literature, then assess each candidate against your inclusion criteria. You learn where extractable SAR exists, how dense it is, and whether the evidence can support the intended dataset.

See if usable SAR exists Read the mining method →

Sources are known; usable rows are not

Data curation

Reduce cleanup, rework, and hidden data errors

Excelra extracts and standardizes structures, targets, activity values, units, and assay context into your chemical entity model. You receive traceable records built for analysis, modeling, and integration, including complex modalities and patent layouts.

See how my schema would look Read the curation method →

The required coverage may already exist

GOSTAR partial backfiles

Put existing curated coverage to work

If GOSTAR already covers your target, modality, disease area, chemotype, or other defined slice, Excelra can license that curated slice with perpetual non-commercial rights to use the delivered data, subject to agreed terms. Your team gets focused historical SAR without rebuilding the same evidence from source.

See if GOSTAR already covers this Read the backfile fit tests →

Get to useful data with less wasted work

Mining vs curation vs GOSTAR partial backfile

Choose the route that answers the immediate need while preserving a path to the final dataset. Screen a corpus before curation, enrich an existing GOSTAR slice, or move customer-supplied sources directly into a defined schema.

Dimension	Data mining	Data curation	GOSTAR partial backfile
Primary job	Determine whether extractable SAR exists, where it sits, and at what yield	Produce traceable, ingestion-ready rows mapped to your fields and ontologies	Acquire an already-curated GOSTAR slice when coverage and schema fit align
Best when	Coverage is unknown; feasibility, competitive intelligence, diligence, or whitespace analysis comes first	You need custom fields, novel ontologies, client sources, or modalities beyond a stock slice	Strong GOSTAR coverage for your slice, and the GOSTAR model is acceptable (or needs only light reshape)
Typical deliverable	Screened source inventory, inclusion disposition, evidence-yield assessment, and next-step recommendation	SDF, CSV, TSV, Excel, JSON, or custom schema with standardized entities, source provenance, and multi-tier QC	Defined SAR export (structures, targets, activities, assays) under agreed rights
Rights posture	Scoped engagement deliverable	Durable internal analytical asset under agreed terms	Perpetual non-commercial rights to use the licensed slice, subject to agreed terms
Pilot to start	Feasibility and source-availability assessment	Representative curation and schema-alignment benchmark	Coverage assessment and representative schema review
You provide	Scientific question; patents vs journals; decision the answer unlocks	Sample patent or field list / schema snippet	Slice definition (target, modality, disease area, chemotype, and so on); required fields; format

The value is choosing the right work in the right order. Use existing GOSTAR coverage where it fits, establish evidence yield before commissioning a larger build, and reserve curation for the fields, relationships, and source material your workflow genuinely requires.

Scope your data requirement Return to solutions

Built and operated by Excelra

Bring the discipline behind GOSTAR to your data requirement

GOSTAR is Excelra’s structure–activity relationship database. It contains data on more than 10.6 million pharmacologically active molecules from over 5 million screened documents, including patent coverage from USPTO, EPO, and WIPO. Excelra developed the data model and curation process and continues to expand and operate the database.

Strengthen automation with scientific ground truth

Public datasets and internal extraction pipelines can accelerate the work, but difficult records still create hidden risk. Excelra supplies validated reference sets and production curation for patent tables, stereochemistry, R-groups, Markush context, and multi-assay layouts that require scientific judgment.

Scientific methods proven at database scale
Mining and curation draw on the same discipline behind GOSTAR: controlled vocabularies, standardized activity types and units, source-linked records, and multi-tier review of structures and assays.
Partial backfiles are GOSTAR slices
License historical SAR for an agreed slice with perpetual non-commercial rights to use the delivered data, subject to the applicable terms.
Scientists adjudicate scientific meaning
Domain experts apply inclusion criteria, resolve chemotype and assay context, and preserve the relationships among structures, targets, activities, and source evidence.
Modalities and deliverables your stack can use
Small molecules, TPDs/degraders, and large molecules when the work calls for them, delivered for on-prem or in-stack use under agreed terms.

Describe your data need Explore GOSTAR slices