Description:
Regression model with 9 QuBiLS-MIDAS descriptors used for the logarithmic values prediction of the minimum effective dose (Log(MED)) in µmol/cm2 to trigger Aedes aegypti repellency.

The training was performed with the Vote meta classifier in Weka 3.9.4 with 10-fold cross-validation, by using the “average” combination rule of these base learners: Gaussian Processes and SMOreg (both with Pearson Universal Kernel (PUK)), and IBk (with K-nearest neighbors = 10 and True cross-validation). The 9 QuBiLS-MIDAS descriptors are namely:

AC[3]_K_F_AB_nCi_2_M11_MP2_T_LGP[2]_c_MID
AC[2]_K_B_AB_nCi_2_M5_NS6_X_LGP[1]_h-s_MID
IB_K_Q_AB_nCi_2_M12_MP0_T_LGL[1-2]_v_MID
AC[3]_S_Q_AB_nCi_2_M12_MP2_T_LGL[1-2]_a_MID
IB_K_B_AB_nCi_2_M3_NS6_n_T_LGP[2]_a-p_MID
IB_S_F_AB_nCi_2_M3_SS4_C_LGL[2-3]_c_MID
IB_S_F_AB_nCi_2_M15_SS1_T_LGP[2]_a_MID
K_Q_AB_nCi_2_M5_NS7_T_LGL[2-3]_c_MID
AC[6]_S_B_AB_nCi_2_M8_NS2_o_T_LGP[2]_m-s_MID

Training set:
71 compounds extracted from 10.1371/journal.pone.0064547

Test set:
8 carboxamides proposed by Oliferenko et al. (10.1371/journal.pone.0064547) were used for external validation.

Performance:
For a 10-fold cross-validation, the statistical parameters (performance without applicability domain) are R = 0.8207, MAE = 0.3714, RMSE = 0.4876, RAE = 58.5363 %, and RRSE = 58.2623 %.

Classification Breakpoint:
The breakpoint is -0.82 µmol/cm2. Values lower than the breakpoint will elicit a repellent response in the Aedes aegypti mosquito. Values greater than or equal to is -0.82 µmol/cm2 represent certain actions occurring, however, these are not enough to activate a repellent reaction in the mosquito.

Reference:
Oliferenko et al. Promising Aedes aegypti Repellent Chemotypes Identified through Integrated QSAR, Virtual Screening, Synthesis, and Bioassay. PLOS ONE. 2013. 8(9): e64547. DOI: 10.1371/journal.pone.0064547