We merge data science with computational chemistry and biology, bioinformatics, cheminformatics and software engineering. Our experience spans biotech and pharmaceutical collaborations, either as a close partner or as an independent analysis provider.
Flexible, custom-tailored engagements, built around your project and your team.
MatGen collaborates either on a per-project basis or by embedding computational experts directly into client teams, full-time or part-time, to run analyses, develop pipelines, and fit into workflows that already exist.
We employ and connect an extensive network of expert data scientists, computational biologists and chemists, and software engineers, which means we can match the expertise to your project rather than the other way round. The people proposed for the work are named before anything is signed.
Our approach is grounded in rigorous data modeling and stated assumptions, producing results that hold up when someone reruns them. Below we set out the stages of discovery and development where that work lands.
Six stages. Most engagements live in one or two of them, and the useful thing about having covered all six is knowing what the next stage will ask of the decision you are making now.
A great deal of current work is deep learning, and deep learning is not appropriate everywhere. A twelve-sample dose–response study is better served by a hierarchical model and an honest confidence interval than by a transformer; a screen across a million compounds is not.
The method therefore follows the data: how much of it there is, how it was collected, what is being asked, and what the consequences are if the answer is wrong. Where a simple model and a complex one perform the same, we use the simple one, since someone will need to read it a year from now.
The same applies to infrastructure. Reproducibility is what keeps an analysis meaningful long after it was first run, so we build it in from the start rather than adding it afterwards. See what that has looked like on real projects.
Whether you need embedded computational expertise, a scoped analytical project, or custom software built around your workflows, the first step is the same. Navigate here to learn how the engagement process with MatGen works.