LASSO selected  (mass_scan)     PLS PCR RR RF EN SPLS Univariate 294_0182 1 1 1 1 1 1 1 187_0439 6 5 2 24 4 3 18 557_1301 2 9 3 5 3 4 4 225_0983 7 3 6 3 5 2 2 622_0122 14 35 4 71 2 0 20 373_1301 3 15 8 9 7 10 9 157_0111 11 2 11 87 6 0 32 1396_0110 18 10 9 88 8 0 38 393_1416  5 6 5 7 9 0 10 843_0095 24 13 13 16 15 0 51 239_0200 23 79 20 93 11 0 23 207_0918 30 32 25 12 16 0 14 499_1599 17 4 14 106 17 0 30 192_0142 13 11 12 35 12 0 26 193_1508 10 75 18 19 10 7 6 738_0133 15 41 7 10 13 0 22 491_0148 26 24 35  161 0 0 144 396_1508 9 36 21 17 14 9 7 323_0837 56 76 53 149 0 0 110 719_0105 20 88 17 90 0 0 35 589_0148 28 33 33 108 0 0 163 303_1245 61 65 43 77 0 0 101 225_1296 21 43 19 62 0 0 64 484_1597 72 155 29 13 0 0 31

Table 2: Ranking of the twenty-four variables (mass_scans) selected by LASSO is shown in the first column. The other columns show the ranks of these twenty-four variables in PLS, PCR, RR, RF, EN, SPLS and univariate regression analysis. In RF ranking was done based on increase in MSE of the OOB samples after permutation of the variable. Out of these twenty-four variables, some are not selected by EN (17 variables selected) and SPLS (10 variables selected) and these are marked as 0. Variables in bold font are also selected in EN and SPLS.