Description:
Regression model with 5 QuBiLS-MIDAS used for the prediction of the Drug Potency (log(1/RD50)) in mg/cm2 to trigger repellent reaction of the mosquito Anopheles gambiae.
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: Linear Regression, IBk (with K-nearest neighbors = 10 and True cross-validation), and Random Forest. The 5 QuBiLS-MIDAS descriptors are namely:
K_TrC_AB_nCi_3_M25(M1)_SS2_T_KA_h_MID
S_TrC_AB_nCi_3_M22(M3)_SS3_T_KA_h_MID
VC_TrB_AB_nCi_3_M21(M8)_SS5_T_LG3L[2-3]_LGL[2-3]_r-p_MID
N2_Tr_AB_nCi_3_M21(M11)_MP3_A_KA_a-e-v_MID
VC_TrQB_AB_nCi_3_M20(M8)_NS6_T_KA_r-p_MID
Training set:
36 compounds extracted from Omolo et al., 2004 10.1016/j.phytochem.2004.08.035
Performance:
For a 10-fold cross-validation, the statistical parameters (performance without applicability domain) are R = 0.8769, MAE = 0.2336, RMSE = 0.3484, RAE = 52.6903 %, and RRSE = 58.4765 %.
Classification Breakpoint:
The breakpoint is 3.41 mg/cm2. Values greater than the breakpoint will elicit a repellent response in mosquito Anopheles gambiae. Values lower or equal to the breakpoint represent certain actions occurring, however, these are not enough to activate a repellent reaction.
Reference:
Omolo et al. Repellency of essential oils of some Kenyan plants against Anopheles gambiae. Phytochemistry. 2004, 65(20), 2797-2802. DOI: 10.1016/j.phytochem.2004.08.035