Our methods
SUPERVISED LEARNING
Methods based on the input information and the final effect and searching for relationships between them (regression and classification methods such as linear and logistic regression, support vector machines, decision trees, Bayesian classifiers, multilayer neural networks, and image analysis).
UNSUPERVISED LEARNING
Methods using only input information (without the final effect), searching for relationships, similarities, and differences between the analyzed objects (e.g. cluster analysis, neural networks and self-organizing maps, principal component analysis, extraction-maximization algorithms).
MOLECULAR DOCKING AND VIRTUAL SCREENING
It allows you to select the best compounds for further in vitro and in vivo testing. The methods used allow both the design of compounds based on the receptor structure (SBDD – structure-based drug design) and the design of compounds based on the structure of the ligand (LBDD – ligand-based drug design).
QSAR
Quantitiative Structure-Activity Relationship
A method that expresses the relationship between chemical structure and properties in a quantitative way, where using information encoded in numbers about a given chemical structure and its toxic effect, a mathematical model can be created, capable of predicting in silico physicochemical properties, biological activity and ADMET – adsorption properties, distribution, metabolism, excretion and toxicity of experimentally untested compounds.
READ-ACROSS
Clustering and read-across assessment methods
A compound similarity approach. According to the guidelines related to the REACH Regulation, one group/category includes substances with similar physicochemical, toxicological and ecotoxicological properties, as well as chemical structure. Thanks to these relationships, it is possible to predict the properties of the target substance using data available for the source compounds. In this way, it is possible to obtain knowledge about new, untested substances only on the basis of their properties and chemical structure.
WoE
Weight of Evidence
The ‘weight of evidence’ procedure refers to collecting, weighing, and evaluating evidence to draw conclusions about the potential toxic effects of a substance. In this process, selected studies are qualified in terms of qualitative acceptability criteria, their compliance with current OECD and other guidelines, GLP status, etc. A weight-of-evidence approach can be performed to assess virtually any in vitro and in vivo toxicity indicator.
