QSAR and QSPR model development
Custom regression and classification models for endpoints where public models fall short — built on curated data, validated internally and externally, and reported with a defined applicability domain.
QSAR Lab builds and validates in silico models that predict the physicochemical properties, toxicity and environmental fate of chemicals and nanomaterials — documented to the standard your regulatory dossier has to meet.
QSAR Lab Sp. z o.o. works at the point where chemistry, statistics and regulation meet. Our clients come to us with a substance and a question that laboratory testing answers slowly, expensively, or not at all: how toxic is this material, how will it behave in water, and can that answer be defended to an authority?
We answer it by building quantitative structure–activity relationship models from the molecular or particle descriptors of the substance itself, then documenting the model so its assumptions, training set and limits are visible to any reviewer. Everything we deliver is written to be read by a third party, not only by us.
The company operates from Gdańsk and works with chemical manufacturers, nanotechnology developers, cosmetics and materials companies, and research consortia across Europe.
You describe the substance and the decision it has to support. We say plainly whether modelling is the right tool.
We assemble the training data, curate it, and calculate the descriptors that characterise your chemistry.
Internal and external validation, applicability domain, and a mechanistic interpretation where one exists.
Predictions plus the documentation an assessor needs, in the format your dossier requires.
Six areas of work. Most projects combine two or three of them, and we are happy to take on a single question as well as a full programme.
Custom regression and classification models for endpoints where public models fall short — built on curated data, validated internally and externally, and reported with a defined applicability domain.
Nano-QSAR modelling for engineered nanoparticles, where size, surface chemistry and coating matter as much as composition. Grouping, read-across and safe-by-design support included.
Predictions prepared as dossier-ready evidence: QMRF and QPRF documentation, justification of non-testing approaches, and help responding to questions from an authority.
When a substance has no data of its own, we identify defensible analogues, quantify the similarity, and write the justification that explains why the transfer of data holds.
Persistence, bioaccumulation and aquatic toxicity estimated in silico, so you can screen a portfolio and put testing budget only where it changes the outcome.
Prediction tools and calculation pipelines built around your workflow, plus hands-on training for R&D and regulatory teams who want to run routine work themselves.
Send a short description of the material and the question you need answered. We reply within two working days with an honest view of whether modelling helps, and what it would cost.