Machine learning applied to prediction of Drug Activity And Target Identification

Welcome TO DrTarget


Doctor target is intended to apply  machine learning technologies to several key stages of the drug discovery process. Based on proprietary chimaeric DB constructions bought together from a number of public sources, such as ChEMBL, NCBI, Uniprot & Open Targets repositories, a number of ad hoc tools have been developed for prediction of pharmacological properties associated with biosystems and molecules.


By using the knowledge of their current profile in the public records, the likelihood of activity for a reasonable number of molecules upon a significant number of targets can be achieved with documented accuracy. Thus, new targets can be identified from phenotypic screening outcomes, and molecule sets that have never touched a biological reagent may be selected for proof of concept at the bench.


These tools can be also employed for the prediction of efficacy in relevant disease systems and  the appropriate assay to assess such properties.


The outcome of the predictive machine learning tools on particular assets can be converted into ad hoc compound collections for screening, Selection of new targets for TV approaches or the design of appropriate assays for assessment of efficacy.


The following tabs on this website will give you some specific examples .
Evaluation and application of machine learning classification and regression algorithms


Machine learning tools: Decision trees, random forests, naive Bayes algorithms, neural networks, support vector machines used to predict compound activity upon assays or targets, reveal protein interactions or involvement in pathways.  Walk the path from disease 2 pathway 2 target 2 molecule.



NCBI, ChemBl and Open Targets
free downloadable info reconfigured
and joined in adHoc databases
optimized for Machine Learning

Any Question?


Sergio Senar,
Data Science
applied to Biomedicine.