Sinopia Biosciences Gets $2.5M NIH Grant for AI Model

Sinopia Biosciences is trying to give earlier laboratory work a longer scientific life. The San Diego biotechnology company announced a $2.5M National Institutes of Health grant on October 20, 2026, to develop and validate an AI model that can help researchers compare metabolomics data collected through different experimental platforms and analytical methods.

The NIGMS SBIR Phase II award funds work on LEADS, Sinopia's drug-discovery platform. Co-founders Aarash Bordbar and Bernhard Palsson built the company around computational biology; principal investigator Bernard Bloem now leads research aimed at making more experimental evidence usable together. For AI drug discovery, the practical opportunity is to learn from measurements that would otherwise remain difficult to compare, while keeping new predictions connected to biological experiments.

What the $2.5M grant funds

Sinopia's October 20 announcement identifies the grant as award R44GM154585 from the National Institute of General Medical Sciences, part of NIH. The project will expand a metabolomics foundation model with proprietary and public datasets, broaden drug perturbation profiling, and evaluate target and compound identification in preclinical disease models. Those activities describe a research program with experimental work attached to the model's development.

Metabolomics measures small molecules that help reveal cellular activity and responses to interventions. Differences between instruments, experimental setups and analytical methods can make results from separate studies difficult to align. Sinopia wants the model to learn biological relationships that remain useful when the source of the measurements changes.

The award also has a documented predecessor. The government's 2024 Phase I record lists $350K for work on large-scale metabolomics data analysis, with Bernard Bloem as principal investigator. That earlier record described the need for standardized methods, computational benchmarking and prospective experiments. The new Phase II announcement extends this line of work; its $2.5M amount represents grant funding, with no equity valuation or venture syndicate announced.

How LEADS connects computation to experiments

Sinopia's LEADS platform, short for LEarn And DiScover, combines multi-omics data, screening, AI and machine learning with mechanistic systems biology. The company's approach examines how biological systems respond to compounds, using those responses to investigate disease mechanisms and potential therapies. Metabolomics supplies one layer of that evidence, alongside other measurements of cellular behavior.

The government's Phase I abstract helps explain why this layer matters. It describes technical and standardization difficulties that have limited the systematic use of proteins and metabolites in data-driven discovery. Its proposed development path joins computational comparisons with prospective experimental validation. A prediction earns practical value when the next experiment can investigate the mechanism it suggests.

For a drug-discovery team, comparability changes the evidence available before another experiment is commissioned. A measurement from a different study could help identify a shared response, distinguish a compound's effects or suggest a question for follow-up. The commercial implication is a larger usable research base, although Sinopia's announcement supplies no quantified reduction in discovery cost or time. Model-transfer performance and experimentally confirmed discoveries will determine how much of that opportunity the project can capture.

The founders' route back into biology

UC San Diego's July research account traces Sinopia to Aarash Bordbar's graduate work in Bernhard Palsson's laboratory. Aarash Bordbar and Bernhard Palsson founded the company in 2014 and established it at the university's Qualcomm Institute Innovation Space. Access to researchers in signaling, chemistry, movement disorders and pharmacokinetics helped put computational work beside the expertise needed to investigate it.

The company's current leadership page lists Aarash Bordbar as CEO, CSO and co-founder, Iman Famili as COO, and Bernhard Palsson as a co-founder and board member. That current structure matters when reading older company announcements, which used different executive titles. The people have stayed connected to the scientific program while their responsibilities have changed.

Sinopia's lead candidate, SB-0110, is a PKA-II modulator in late-stage preclinical development for Parkinson's disease. The university account describes published research in rodent and nonhuman-primate models, including investigation of levodopa benefit and dyskinesia. Those are animal findings, with human benefit still awaiting clinical study. For the foundation-model project, the existing program offers company-specific context: Sinopia has already taken computationally informed discovery into a drug-development effort, and is funding work intended to broaden the evidence behind future programs.

Where a pharmaceutical partner enters the picture

Sinopia has another route for putting its discovery work to use. Its March collaboration with Ono Pharmaceutical applies the company's metabolomics approach to identifying and validating targets for an undisclosed set of rare metabolic disorders. The agreement includes research funding, an upfront payment and eligibility for future milestones, with financial terms undisclosed.

Ono may advance targets identified through the collaboration toward new medicines. That arrangement gives the platform an external research relationship alongside Sinopia's internal pipeline. It also makes the eventual handoff concrete: findings have to be useful to another organization deciding which biological mechanisms deserve development work. The collaboration is separate from the NIH grant and does not make Ono a named investor in this award.

Sinopia's pipeline overview identifies neurology, oncology and genetic diseases as its initial therapeutic areas. Reusing evidence across experimental contexts could support that breadth, provided the model can preserve the relationships that matter in each setting. The $2.5M project places that question inside a funded development and validation program. As Sinopia adds measurements from more sources, its researchers will be working out which findings can travel with them into the next disease model and, eventually, another drug team's decisions.