Mycobacteria represent a group of pathogens which cause serious diseases in mammals, including the lethal tuberculosis
(Mycobacterium tuberculosis). Despite the mortality of this community-acquired and nosocomial disease mentioned
above, other mycobacteria may cause similar infections, acting as dangerous opportunistic pathogens. Additionally,
resistant strains belonging to Mycobacterium spp. have emerged. Thus, the design of novel antimycobacterial agents is a
challenge for the scientific community. In this sense, chemoinformatics has played a vital role in drug discovery, helping
to rationalize chemical synthesis, as well as the evaluation of pharmacological and ADMET (absorption, distribution, metabolism,
excretion, toxicity) profiles in both medicinal and pharmaceutical chemistry. Until now, there is no in silico
methodology able to assess antimycobacterial activity and ADMET properties at the same time. This work introduces the
first multitasking model based on quantitative-structure biological effect relationships (mtk-QSBER) for simultaneous
prediction of antimycobacterial activities and ADMET profiles of drugs/chemicals under diverse experimental conditions.
The mtk-QSBER model was constructed by using a large and heterogeneous dataset of compounds (more than 34600
cases), displaying accuracies higher than 90% in both, training and prediction sets. To illustrate the utility of the present
model, several molecular fragments were selected and their contributions to different biological effects were calculated
and analyzed. Also, many properties of the investigational drug TMC-207 were predicted. Results confirmed that, from
one side, TMC-207 can be a promising antimycobacterial drug, and on the other hand, this study demonstrates that the
present mtk-QSBER model can be used for virtual screening of safer antimycobacterial agents.