Prioris.ai

Prioris.ai is a Canadian life science consultancy and analytics firm headquartered in Ottawa, Ontario. The company specialises in probabilistic predictive modelling — often described as Bayesian machine learning, probabilistic programming, or model-based machine learning — and applies these methods to problems across drug discovery, preclinical safety, and clinical development. Rather than offering a generic software platform, Prioris.ai positions itself as a bespoke modelling partner: it builds custom, interpretable predictive models for clients in the life science industry. The firm's service portfolio spans five broad areas. First, predictive modelling for safety pharmacology, toxicology, compound ranking and prioritisation, and assay optimisation. Second, Bayesian statistics for statistical inference in clinical and preclinical experiments. Third, design of experiments (DoE) to maximise information gained from experiments, avoid confounding, and produce generalisable results. Fourth, automation of analytical workflows to save time, reduce error, and increase reproducibility. Fifth, regulatory writing, with a focus on clinical regulatory writing and submissions. Prioris.ai's technical philosophy centres on probabilistic predictive modelling. The company argues that this approach deals with uncertainty in a principled way — uncertainty in data, parameters, and models is propagated and reported for all predictions, avoiding over-confident outputs. It also allows external information, such as scientific knowledge, databases, experiments, and expert judgement, to be incorporated into model structure. Because models are constructed using background scientific knowledge, they remain interpretable rather than opaque. Finally, the approach is designed to work with small data, a common constraint in life science research where large sample sizes are often unavailable. A flagship example of the company's work is its prediction model for drug-induced liver injury (DILI). Prioris.ai states that this was the first fully Bayesian DILI prediction model, published in Chemical Research in Toxicology, with a preprint also made available. The model has reportedly been in production since the end of 2017 and was short-listed for an AstraZeneca Science Award. Supplementary material accompanying the publication includes code for R/Stan and Python/PyMC3, reflecting the company's use of open probabilistic programming tools. Prioris.ai operates as a private company and serves clients in the life science sector, including pharmaceutical and biotechnology organisations. Its combination of statistical rigour, domain-specific scientific knowledge, and regulatory awareness positions it at the intersection of data science and drug development. The company's public presence is centred on its website, which outlines services, methodology, and featured work.

Prioris.ai is a Canadian company based in Ottawa, Ontario, operating in the life science analytics and consulting space. Public information about the firm is relatively limited, but its website and Wikidata entry establish it as a company in Ottawa, Canada, focused on predictive modelling for the life science industry. The company's most documented milestone is the development of a fully Bayesian model for predicting drug-induced liver injury (DILI). According to Prioris.ai, this was the first fully Bayesian DILI prediction model and was published in the journal Chemical Research in Toxicology, with a preprint made available online. The company states that the model has been in production since the end of 2017, indicating that Prioris.ai was active and delivering applied modelling work by that time. The model was also short-listed for an AstraZeneca Science Award, suggesting recognition within the pharmaceutical industry. The supplementary material for the DILI publication includes code for R/Stan and Python/PyMC3, which reflects the company's technical stack and its commitment to reproducible, open probabilistic programming. This aligns with the firm's stated methodology: probabilistic predictive modelling that propagates uncertainty, incorporates external scientific knowledge, remains interpretable, and works with small datasets. Over time, Prioris.ai has broadened its service offering beyond a single model. Its website lists five service areas: predictive modelling (safety pharmacology, toxicology, compound ranking/prioritisation, assay optimisation), Bayesian statistics for clinical and preclinical experiments, design of experiments, automation of analytical workflows, and regulatory writing for clinical submissions. This breadth suggests an evolution from a specialised modelling project toward a fuller consultancy serving multiple stages of drug development. The company operates as a private entity and does not appear to be listed on a public stock exchange. Its client base is understood to be within the life science industry, including pharmaceutical and biotechnology organisations that require bespoke statistical and modelling expertise. As of the available evidence, Prioris.ai remains active, with its website serving as the primary public source of information about its services and featured work.

• 2017

  Fully Bayesian DILI prediction model reportedly enters production

• DILI prediction model published in Chemical Research in Toxicology (preprint available)

• DILI model short-listed for an AstraZeneca Science Award

Bespoke, interpretable, probabilistic (Bayesian) predictive modelling partner for the life science industry, emphasising uncertainty quantification, small-data capability, and regulatory-grade rigour.

• Fully Bayesian DILI prediction model

  First fully Bayesian drug-induced liver injury prediction model, published in Chemical Research in Toxicology, in production since end 2017, short-listed for an AstraZeneca Science Award. Supplementary material includes R/Stan and Python/PyMC3 code.

• Predictive modelling services

  Safety pharmacology, toxicology, compound ranking/prioritisation, and assay optimisation.

• Bayesian statistics services

  Statistical inference for clinical and pre-clinical experiments, mainly from a Bayesian perspective.

• Design of Experiments (DoE)

  Maximising information gained from experiments, avoiding confounding, and obtaining generalisable results.

• Analytical workflow automation

  Save time, reduce error, increase reproducibility.

• Regulatory writing

  Specialising in clinical regulatory writing and submissions.

1. 1Prioris.ai Inc. When predictions matterprioris.ai
2. 2Prioris.ai (Canada) - Wikidatawikidata.org