Description:
Regression model of sensillum SST-C with 7 QuBiLS-MIDAS descriptors used for prediction of spike amplitude of the extracellularly record of ORN response potentials of the antennal trichoid sensilla to Cx. quinquefasciatus repellents.

The training was performed with Additive Regression algorithm as default in Weka 3.9.4 with a 10-fold cross-validation using SMOreg (with Pearson Universal Kernel (PUK)) as classifier. The 7 QuBiLS-MIDAS descriptors are namely:

IB_S_F_AB_nCi_2_M11_SS1_T_LGP[2]_c_MID
N2_TrQB_AB_nCi_3_M19(M11)_MP6_T_LGBA[1.570-1.884]_r-p_MID
AC[2]_S_F_AB_nCi_2_M16_MP1_T_LGP[1]_c_MID
S_B_AB_nCi_2_M13_NS4_T_LGL[5-6]_a-s_MID
S_B_AB_nCi_2_M10_SS7_T_LGL[5-6]_c-h_MID
AC[5]_S_Q_AB_nCi_2_M12_MP0_T_LGL[1-2]_a_MID
AC[5]_S_F_AB_nCi_2_M8_SS3_T_LGP[1]_v_MID

Training set:
50 compounds extracted from the Liu, et al., 2013 original experimental research constitute the training set 10.1016/j.jinsphys.2013.08.016

Performance:
For a 10-fold cross-validation, the statistical parameters (performance without applicability domain) are R = 0.9133, MAE = 6.1301, RMSE = 7.5744, RAE = 44.0349 %, and RRSE = 43.0061 %.

Classification Breakpoint:
The breakpoint for sensillum SST-C is 13.5 spikes/s. Values greater than or equal to the breakpoint will elicit a response in this specific sensillum. Lower values represent certain actions occurring, however, they are not enough to activate the sensillum response.

Reference:
Liu et al. Olfactory responses of the antennal trichoid sensilla to chemical repellents in the mosquito, Culex quinquefasciatus. 2013, 59(11), 1169-1177. DOI: 10.1016/j.jinsphys.2013.08.016