S. No. Descriptor Descriptor class Physical meaning Average regression coefficient (incidence) 1 S3K Topological 3-Path Kier alpha-modified shape index 2.135 (10) 2 CIC1 Topological Complementary information content index of 1-order neighborhood symmetry -1.949 (4) 3 SEigZ Topological Eigenvalue sum from Z weighted distance matrix (Barysz matrix) 1.733 (4) 4 SEigm Topological Eigenvalue sum from mass weighted distance matrix 1.811 (4) 5 BELv6 BCUT Lowest eigenvalue n. 6 of Burden matrix / weighted by atomic van der Waals volumes 1.861 (7) 6 BELe1 BCUT Lowest eigenvalue n. 1 of Burden matrix / weighted by atomic Sanderson electronegativities -0.854 (4) 7 BELp6 BCUT Lowest eigenvalue n. 6 of Burden matrix / weighted by atomic polarizabilities 1.382 (5) 8 GGI3 Galvez topological charge indices 3-Order topological charge index -2.524 (12) 9 ATS5m 2D-autocorrelation Broto-Moreau autocorrelation of a topological structure - lag 5 / weighted by atomic masses 1.765 (1) 10 MATS4m 2D-autocorrelation Moran autocorrelation - lag 4 / weighted by atomic masses 0.747 (1) 11 nCs Functional number of total secondary C(sp3) 1.379 (1) 12 nNR2 Functional number of tertiary amines (aliphatic) -0.741 (2) 13 C-006 Atom centered fragment CH2RX 1.198 (2) 14 C-025 Atom centered fragment R--CR--R -1.161 (7) 15 N-075 Atom centered fragment R--N--R / R--N--X 0.668 (1) aThe descriptors are identified from the five parameter models, emerged from CP-MLR protocol with filter-1 as 0.79, filter-2 as 2.0, filter-3 as 0.837, and filter-4 as 0.3 ≤ q2≤1.0 with a training set of 40 compounds. bThe average regression coefficient of the descriptor corresponding to all models and the total number of its incidence. The arithmetic sign of the coefficient represents the actual sign of the regression coefficient in the models

Table 2: Identified descriptorsa along with their physical meaning, average regression coefficient and incidenceb, in modeling the MMP-13 inhibition activity.