In silico evaluation of pharmacokinetic properties and molecular docking for the identification of potential anticancer compounds.

INTRODUCTION: In the pharmaceutical field, the rapid and accurate characterization of physicochemical properties is essential for drug development. In this context, in silico methodologies facilitate the early-stage prediction of ADME (Absorption, Distribution, Metabolism, and Excretion) and toxicity (ADME-Tox) parameters, reducing experimental costs and accelerating the screening of viable drug candidates. Computational approaches such as QSAR, SAR, and QSPR enable the assessment of biological activity and pharmacokinetic behavior. OBJECTIVE: This study aimed to evaluate the ADME(T) profiles of 58 organic compounds using computational tools, establish predictive models for toxicity, and assess inhibitory potential against the TLK2 kinase domain (PDB: 5O0Y)-a protein implicated in breast cancer and intellectual disability. METHODOLOGY: Chemical structures were optimized using the MMFF94 force field. ADME-Tox descriptors-including Log P, Log S, Caco-2 permeability, CYP450 interactions, hERG inhibition, LD50, and DILI-were calculated using SwissADME and PreADMET. Data analysis included Pearson correlation, PCA, hierarchical clustering, and construction of a cosine similarity network. A Random Forest regression model was implemented to predict LD₅₀ values, and molecular docking simulations were conducted using PyRx and Discovery Studio. RESULTS: Correlation and PCA analyses revealed key trends, including a strong relationship between Log P and Log D, and groupings based on structural similarity. The Random Forest model demonstrated strong predictive performance for LD50 (r2 = 0.8410; RMSE = 0.1112), with five-fold cross-validation confirming robustness. Molecular docking identified several compounds with favorable binding affinities to TLK2, notably compound CC-43, which showed the strongest interaction (-8.2 kcal/mol) and a moderate theoretical toxicity profile (LD50 = 3.186). CONCLUSIONS: The integration of ADME(T) profiling, machine learning, and molecular docking provides a comprehensive and reproducible computational strategy for drug discovery. The approach enabled the identification of compounds with favorable pharmacokinetic properties and selective inhibitory potential, supporting compound CC-43 as a promising candidate for further exploration in breast cancer therapeutics.