Platform
Inference was built by statistical geneticists who are also software engineers. Teams who run their own population-scale studies and discover their own drug targets. Designed from the ground up with AI in mind, every feature is accessible via API and optimized for LLM integration. It powers our internal pipeline and pharma collaborations—real-world programs, not tech demos.
Other platforms give you static reports and pre-run analyses. We took a fundamentally different approach.
Inference provides harmonized data that's instantly available for on-demand downstream analyses. Query any gene and get results across hundreds of millions of data points in seconds. Run colocalization or Mendelian randomization. Filter and combine studies in real time. See the underlying data behind every result, and ask the next question immediately: no waiting for consultants or custom analysis packages.
The Inference platform reflects how discovery actually happens: iteratively, dynamically, following the data wherever it leads.
Yeah, us too. That's why we built tools that enable real-time querying of full, variant-level summary stats for every loaded GWAS and QTL study.
Stop wondering if incomplete data is hiding an important hit just below genome-wide significance. Instantly see all molecular associations that could reveal mechanism or pinpoint causal genes.
Inference includes interactive tools built to let scientists dynamically explore data across massive datasets.
Query by gene and see full GWAS and QTL summary stats side-by-side in a single unified view. Quick access to expression levels, differential quantification in disease, drug info, and external links. Launch analyses and visualizations with a click.
Interactive locus plots that render from live summary stats streaming from the API, not cached images. Dynamically add both GWAS and QTL studies, change LD references, and filter based on variant annotations.
Construct complex queries that run across all datasets. Filter by fields including phenotype, data type, variant effect, p-value, and finemap probability. Or do a deep dive into all the results from your favorite GWAS, exploring it locus-by-locus.
Inference was designed from the ground up to map the molecular mechanisms underlying genetic associations by integrating across multi-omic data. No guesswork about how data types connect. GWAS are colocalized with molecular QTLs to identify causal genes and generate therapeutic hypotheses. Mendelian randomization with genome-wide instruments is used to further refine hypotheses.
By anchoring analyses on the genome, Inference establishes causation, not just correlation.
Developer Access
A proprietary API with hundreds of endpoints powers the entire platform. Every Explorer, every agent, every feature, all running on the same infrastructure. This isn't a dashboard bolted onto a database. It's a unified system designed for programmatic access at scale.
Learn how Inference Agents can execute PhD-level genomic analyses autonomously.
Explore AI Agents →