Two novel series of potent and selective matrix metalloproteinase (MMP) inhibitors, involving unique binding
mode at the active site and not interacting with the catalytic zinc, were collectively investigated to quantify their inhibition
actions in relation with chemometric descriptors. Significant correlations, between their MMP13 inhibition activity and
2D-descriptors, were obtained through the combinatorial protocol in multiple linear regression (CP-MLR) computational
procedure. The derived bi-variant models, validated internally and externally, were able to account for 86.5% of variance
in the observed MMP13 inhibition activities. The filtered descriptors, from CP-MLR, satisfactorily explained the biological
phenomenon under investigation. However the descriptors, MPC10, N-075 and C-030 accounting, respectively, for
molecular path (bond) count of order 10, the structural fragments CH--N--CH and N--CH--N of aromatic ring remained
prime to address the MMP13 inhibition actions of the compounds. From the highest significant correlations, it appeared
that the higher value of MPC10 and absence of these aromatic ring fragments, are conducive in further improvement of
MMP13 inhibition activity of a compound. The partial least squares (PLS) analysis has further confirmed the dominance
of the identified descriptors. In the analysis, two-components remained optimum for these descriptors which are able to
explain 89.3% of variances. Applicability domain (AD) analysis revealed that the suggested models have acceptable predictability.
All the compounds remained within the AD of the proposed models and were evaluated correctly. Based on
the inferences drawn from the study, some new analogues were suggested for further exploration. Their predicted MMP13
inhibition activities were much higher than the highest active congener of the original series.