Results of FLU Analysis ======================= In this analysis we are using the data from all 17 subjects (no subjects disqualified). The classification that we are trying to learn and predict is the following: Subjects 1,5,6,7,8,10,12 are considered symptomatic and all others are considered asymptomatic. For the temporal abstraction, we define baseline, early, middle, and late intervals are -24...0, 5...21.5, 29...53, and 60...108, respectively. Results for Composite Single-Gene Predictors -------------------------------------------- Early Predictor: The best ternary logic predictor synthesized, which consists in average of 2.29 single gene-predictors, makes prediction in 13 out of 17 cases with 5 errors (which directly translates into an error rate of 38%). The normalized weighted error rate (based on an equivalent binary predictor) is 44%. The p-value determined using 100 randomization tests is ~ 0.45. The best predictor based on all 17 subjects (i.e. without crossvalidation) contains exactly the following two genes: NOL5A and UCP1 (both positive predictors). Middle Predictor: The best predictor consists in average of 2.12 genes and makes predictions in 11 out of 17 cases with 1 errors (yielding a direct error rate of 9%). The normalized error rate is 20% and the p-value is ~ 0.01. The best predictor based on all 17 subjects consists of three genes: C18orf25, STYXL1 (both positive), and VPREB3 (negative). Late Predictor: The best predictor consists in average of 1.94 genes and makes predictions in 15 out of 17 cases with 3 errors (translating into a direct error rate of 20%). The normalized error rate is 23% with a p-value of ~ 0.06. The best predictor based on all 17 subjects consists of two genes: ELMO2 (positive), ZNF91 (negative), IFI27 (positive). Relevant Log-File Fragments Testing top composite predictors at time 13.25 ... Average Best Composite Predictor: Top-2.29 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 13 (0.76) Err: 5 (0.38) NErr: 0.41 WErr: 0.42 NWErr: 0.44 CScore: 0.44 Ok: 1,2,4,5,9,11,12,14 Err: 6,7,8,10,15 CompPred Top-1 - Sub: 17 Pred: 2 (0.12) Err: 1 (0.50) NErr: 0.50 WErr: 0.59 NWErr: 0.51 CScore: 0.51 Ok: 9 Err: 6 CompPred Top-2 - Sub: 17 Pred: 11 (0.65) Err: 3 (0.27) NErr: 0.35 WErr: 0.29 NWErr: 0.36 CScore: 0.36 Ok: 1,2,4,5,9,11,12,14 Err: 6,7,15 CompPred Top-3 - Sub: 17 Pred: 14 (0.82) Err: 5 (0.36) NErr: 0.38 WErr: 0.38 NWErr: 0.40 CScore: 0.40 Ok: 1,2,3,4,5,9,12,13,14 Err: 6,7,8,15,17 CompPred Top-4 - Sub: 12 Pred: 12 (1.00) Err: 3 (0.25) NErr: 0.25 WErr: 0.26 NWErr: 0.34 CScore: 0.34 Ok: 2,3,4,5,9,11,12,13,14 Err: 10,16,17 Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.01 p(e <= 0.10) = 0.02 p(e <= 0.15) = 0.03 p(e <= 0.20) = 0.03 p(e <= 0.25) = 0.08 p(e <= 0.30) = 0.17 p(e <= 0.35) = 0.22 p(e <= 0.40) = 0.34 p(e <= 0.45) = 0.45 p(e <= 0.50) = 0.55 Computing top composite predictors at time 13.25 (without crossvalidation) ... Computing predictors at time 13.25 ... Defining composite predictors ... 1. Pred NOL5A - Pat: Pos - Sub: 17 Pred: 8 (0.47) Err: 0 (0.00) NErr: 0.26 WErr: 0.00 NWErr: 0.27 Ok: 2,6,7,9,11,12,14,15 Err: 2. Pred UCP1 - Pat: Pos - Sub: 17 Pred: 7 (0.41) Err: 0 (0.00) NErr: 0.29 WErr: 0.00 NWErr: 0.28 Ok: 1,4,5,6,7,9,15 Err: 3. Pred SLAMF7 - Pat: Pos - Sub: 17 Pred: 9 (0.53) Err: 3 (0.33) NErr: 0.41 WErr: 0.33 NWErr: 0.41 Ok: 1,3,4,6,9,13 Err: 8,11,17 4. Pred NDUFA1 - Pat: Pos - Sub: 17 Pred: 7 (0.41) Err: 3 (0.43) NErr: 0.47 WErr: 0.41 NWErr: 0.46 Ok: 2,7,8,11 Err: 10,16,17 Testing top composite predictors at time 13.25 ... Best Composite Predictor: Top-2 Score: 0.17 Testing top composite predictors at time 41.0 ... Average Best Composite Predictor: Top-2.12 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 11 (0.65) Err: 1 (0.09) NErr: 0.24 WErr: 0.07 NWErr: 0.20 CScore: 0.20 Ok: 1,3,5,6,7,8,10,12,15,17 Err: 2 CompPred Top-1 - Sub: 17 Pred: 10 (0.59) Err: 1 (0.10) NErr: 0.26 WErr: 0.08 NWErr: 0.23 CScore: 0.23 Ok: 1,3,5,6,7,8,10,12,17 Err: 2 CompPred Top-2 - Sub: 17 Pred: 13 (0.76) Err: 3 (0.23) NErr: 0.29 WErr: 0.19 NWErr: 0.25 CScore: 0.25 Ok: 1,3,5,6,7,8,10,12,14,17 Err: 2,11,16 CompPred Top-3 - Sub: 10 Pred: 10 (1.00) Err: 1 (0.10) NErr: 0.10 WErr: 0.08 NWErr: 0.23 CScore: 0.23 Ok: 1,3,5,6,7,8,10,12,15 Err: 2 Randomization Test at time 41.0 ... Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.00 p(e <= 0.25) = 0.05 p(e <= 0.30) = 0.11 p(e <= 0.35) = 0.16 p(e <= 0.40) = 0.30 p(e <= 0.45) = 0.42 p(e <= 0.50) = 0.63 Composite Predictor Top-1 - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.05 p(e <= 0.25) = 0.10 p(e <= 0.30) = 0.18 p(e <= 0.35) = 0.25 p(e <= 0.40) = 0.35 p(e <= 0.45) = 0.41 p(e <= 0.50) = 0.54 Composite Predictor Top-2 - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.00 p(e <= 0.25) = 0.03 p(e <= 0.30) = 0.08 p(e <= 0.35) = 0.17 p(e <= 0.40) = 0.27 p(e <= 0.45) = 0.43 p(e <= 0.50) = 0.57 Computing top composite predictors at time 41.0 (without crossvalidation) ... Computing predictors at time 41.0 ... Defining composite predictors ... 1. Pred C18orf25 - Pat: Pos - Sub: 17 Pred: 13 (0.76) Err: 1 (0.08) NErr: 0.18 WErr: 0.06 NWErr: 0.15 Ok: 1,3,5,6,7,8,9,10,11,12,14,17 Err: 2 2. Pred STYXL1 - Pat: Pos - Sub: 17 Pred: 11 (0.65) Err: 1 (0.09) NErr: 0.24 WErr: 0.07 NWErr: 0.21 Ok: 1,5,6,7,8,11,12,14,16,17 Err: 9 3. Pred VPREB3 - Pat: Neg - Sub: 17 Pred: 9 (0.53) Err: 1 (0.11) NErr: 0.29 WErr: 0.13 NWErr: 0.30 Ok: 1,3,4,7,9,12,15,16 Err: 8 Testing top composite predictors at time 41.0 ... Best Composite Predictor: Top-3 Score: 0.07 Testing top composite predictors at time 84.0 ... Average Best Composite Predictor: Top-1.94 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 15 (0.88) Err: 3 (0.20) NErr: 0.24 WErr: 0.20 NWErr: 0.23 CScore: 0.23 Ok: 2,3,4,5,6,7,9,10,12,14,16,17 Err: 8,13,15 CompPred Top-1 - Sub: 17 Pred: 14 (0.82) Err: 2 (0.14) NErr: 0.21 WErr: 0.12 NWErr: 0.20 CScore: 0.20 Ok: 2,3,4,5,6,7,9,10,12,14,16,17 Err: 13,15 CompPred Top-2 - Sub: 16 Pred: 14 (0.88) Err: 2 (0.14) NErr: 0.19 WErr: 0.12 NWErr: 0.20 CScore: 0.20 Ok: 2,3,4,5,6,7,9,10,12,14,16,17 Err: 13,15 CompPred Top-3 - Sub: 9 Pred: 9 (1.00) Err: 1 (0.11) NErr: 0.11 WErr: 0.14 NWErr: 0.33 CScore: 0.33 Ok: 1,2,3,4,9,14,16,17 Err: 8 Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.01 p(e <= 0.20) = 0.02 p(e <= 0.25) = 0.06 p(e <= 0.30) = 0.12 p(e <= 0.35) = 0.16 p(e <= 0.40) = 0.29 p(e <= 0.45) = 0.39 p(e <= 0.50) = 0.53 Computing top composite predictors at time 84.0 (without crossvalidation) ... Computing predictors at time 84.0 ... Defining composite predictors ... 1. Pred ELMO2 - Pat: Pos - Sub: 17 Pred: 13 (0.76) Err: 0 (0.00) NErr: 0.12 WErr: 0.00 NWErr: 0.12 Ok: 2,3,4,5,6,7,9,10,11,12,14,16,17 Err: 2. Pred ZNF91 - Pat: Neg - Sub: 17 Pred: 11 (0.65) Err: 0 (0.00) NErr: 0.18 WErr: 0.00 NWErr: 0.20 Ok: 2,3,4,7,9,11,12,14,15,16,17 Err: 3. Pred IFI27 - Pat: Pos - Sub: 17 Pred: 15 (0.88) Err: 7 (0.47) NErr: 0.47 WErr: 0.39 NWErr: 0.40 Ok: 1,5,6,7,8,10,12,17 Err: 2,4,9,13,14,15,16 Testing top composite predictors at time 84.0 ... Best Composite Predictor: Top-3 Score: 0.07 Results for Composite Single-Gene Predictors with Preselected Genes ------------------------------------------------------------------- Early Predictor: The best ternary logic predictor synthesized, which consists in average of 1.94 single gene-predictors, makes prediction in 14 out of 17 cases with 5 errors (which directly translates into an error rate of 36%). The normalized weighted error rate (based on an equivalent binary predictor) is 41%. The p-value determined using 100 randomization tests is ~ 0.26. The best predictor based on all 17 subjects (i.e. without crossvalidation) contains exactly the following two genes: FADD and FCER1G (both positive predictors). Middle Predictor: The best predictor consists in average of 2.00 genes and makes predictions in 12 out of 17 cases with 3 errors (yielding a direct error rate of 25%). The normalized error rate is 30% and the p-value is ~ 0.05. The best predictor based on all 17 subjects consists of two genes: MYD88 and IFIT2 (both positive). Late Predictor: The best predictor consists in average of 1.06 genes and makes predictions in 14 out of 17 cases with 2 errors (translating into a direct error rate of 14%). The normalized error rate is 17% with a p-value of ~ 0.02. The best predictor based on all 17 subjects consists of two genes: FCER1G and MX1 (both positive). Relevant Log-File Fragments Testing top composite predictors at time 13.25 ... Average Best Composite Predictor: Top-1.94 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 14 (0.82) Err: 5 (0.36) NErr: 0.38 WErr: 0.39 NWErr: 0.41 CScore: 0.41 Ok: 1,3,7,8,11,13,14,15,16 Err: 5,6,9,10,12 CompPred Top-1 - Sub: 17 Pred: 7 (0.41) Err: 5 (0.71) NErr: 0.59 WErr: 0.77 NWErr: 0.62 CScore: 0.62 Ok: 13,16 Err: 5,6,9,10,12 CompPred Top-2 - Sub: 17 Pred: 14 (0.82) Err: 5 (0.36) NErr: 0.38 WErr: 0.39 NWErr: 0.41 CScore: 0.41 Ok: 1,3,7,8,11,13,14,15,16 Err: 5,6,9,10,12 Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.00 p(e <= 0.25) = 0.02 p(e <= 0.30) = 0.07 p(e <= 0.35) = 0.09 p(e <= 0.40) = 0.22 p(e <= 0.45) = 0.39 p(e <= 0.50) = 0.48 Computing top composite predictors at time 13.25 (without crossvalidation) ... Computing predictors at time 13.25 ... Defining composite predictors ... 1. Pred FADD - Pat: Pos - Sub: 17 Pred: 9 (0.53) Err: 2 (0.22) NErr: 0.35 WErr: 0.23 NWErr: 0.35 Ok: 1,8,12,13,14,15,16 Err: 6,9 2. Pred FCER1G - Pat: Pos - Sub: 17 Pred: 7 (0.41) Err: 1 (0.14) NErr: 0.35 WErr: 0.17 NWErr: 0.36 Ok: 3,7,11,12,13,16 Err: 10 Testing top composite predictors at time 13.25 ... Best Composite Predictor: Top-2 Score: 0.30 Testing top composite predictors at time 41.0 ... Average Best Composite Predictor: Top-2.00 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 12 (0.71) Err: 3 (0.25) NErr: 0.32 WErr: 0.23 NWErr: 0.30 CScore: 0.30 Ok: 1,5,6,7,8,12,15,16,17 Err: 2,10,13 CompPred Top-1 - Sub: 17 Pred: 13 (0.76) Err: 4 (0.31) NErr: 0.35 WErr: 0.28 NWErr: 0.32 CScore: 0.32 Ok: 1,5,6,7,8,12,15,16,17 Err: 2,10,11,13 CompPred Top-2 - Sub: 17 Pred: 12 (0.71) Err: 3 (0.25) NErr: 0.32 WErr: 0.23 NWErr: 0.30 CScore: 0.30 Ok: 1,5,6,7,8,12,15,16,17 Err: 2,10,13 Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.00 p(e <= 0.25) = 0.00 p(e <= 0.30) = 0.05 p(e <= 0.35) = 0.07 p(e <= 0.40) = 0.22 p(e <= 0.45) = 0.40 p(e <= 0.50) = 0.57 Computing top composite predictors at time 41.0 (without crossvalidation) ... Computing predictors at time 41.0 ... Defining composite predictors ... 1. Pred MYD88 - Pat: Pos - Sub: 17 Pred: 10 (0.59) Err: 0 (0.00) NErr: 0.21 WErr: 0.00 NWErr: 0.19 Ok: 1,5,6,7,8,11,12,15,16,17 Err: 2. Pred IFIT2 - Pat: Pos - Sub: 17 Pred: 13 (0.76) Err: 2 (0.15) NErr: 0.24 WErr: 0.12 NWErr: 0.20 Ok: 1,4,5,6,7,8,10,12,15,16,17 Err: 2,13 Testing top composite predictors at time 41.0 ... Best Composite Predictor: Top-2 Score: 0.17 Testing top composite predictors at time 84.0 ... Average Best Composite Predictor: Top-1.06 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 14 (0.82) Err: 2 (0.14) NErr: 0.21 WErr: 0.12 NWErr: 0.17 CScore: 0.17 Ok: 1,3,4,5,6,7,8,10,12,14,16,17 Err: 13,15 CompPred Top-1 - Sub: 17 Pred: 15 (0.88) Err: 2 (0.13) NErr: 0.18 WErr: 0.11 NWErr: 0.15 CScore: 0.15 Ok: 1,3,4,5,6,7,8,10,11,12,14,16,17 Err: 13,15 CompPred Top-2 - Sub: 17 Pred: 16 (0.94) Err: 3 (0.19) NErr: 0.21 WErr: 0.16 NWErr: 0.18 CScore: 0.18 Ok: 1,3,4,5,6,7,8,9,10,12,14,16,17 Err: 2,13,15 Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.01 p(e <= 0.20) = 0.02 p(e <= 0.25) = 0.04 p(e <= 0.30) = 0.15 p(e <= 0.35) = 0.19 p(e <= 0.40) = 0.30 p(e <= 0.45) = 0.54 p(e <= 0.50) = 0.65 Computing top composite predictors at time 84.0 (without crossvalidation) ... Computing predictors at time 84.0 ... Defining composite predictors ... 1. Pred FCER1G - Pat: Pos - Sub: 17 Pred: 15 (0.88) Err: 2 (0.13) NErr: 0.18 WErr: 0.11 NWErr: 0.15 Ok: 1,3,4,5,6,7,8,10,11,12,14,16,17 Err: 13,15 2. Pred MX1 - Pat: Pos - Sub: 17 Pred: 15 (0.88) Err: 4 (0.27) NErr: 0.29 WErr: 0.22 NWErr: 0.25 Ok: 1,4,5,6,7,8,9,10,12,16,17 Err: 2,11,13,15 Testing top composite predictors at time 84.0 ... Best Composite Predictor: Top-1 Score: 0.15 Preliminary Results for Composite Cluster Predictors ---------------------------------------------------- Early Predictors: The best computed ternary predictor consists in average of 3.41 (sub)clusters. It makes predictions in 9 out of 17 cases with 5 errors (translating into 56% error rate). The normalized error rate is 51%. The best predictor based on all 20 subjects consists of 2 (sub)clusters named 13.0 and 15.0 (both positive). See table below for the corresponding sets of genes. Middle Predictors: The best ternary predictor consists in average of 2.88 (sub)clusters. It makes predictions in 13 out of 17 cases with 1 error (yielding an error rate of 8%). The normalized error rate is 15% and the p-value is ~ 0.02. The best predictor based on all 20 subjcts has 3 (sub)clusters: 3.65, 15.0, 8.0 (all positive). Late Predictors: The best ternary predictor consists in average of 2.06 (sub)clusters. It makes predictions in 16 out of 17 cases with 2 errors (yielding an error rate of 12%). The normalized error rate is 12% with a p-value of ~ 0.01. The best predictor based on all 20 subjects has 2 (sub)clusters: 8.0 (positive) and 1.13 (negative). Table of Relevant Clusters subcluster 13.0: { RPL37A RPL38 } subcluster 15.0: { FLI1 VAMP3 } subcluster 3.65: { CD46 PDE4B BACH1 } subcluster 8.0: { ZFYVE26 KIAA0226 } subcluster 1.13: { PHB2 FBL AKR1B1 23157-at LAT BIN1 TRBC1 RANGRF IMPDH2 LCK MAL CUTA TMEM204 SKAP1 PABPC4 CD2 FAIM3 CD27 } Relevant Log-File Fragments Testing top composite predictors at time 13.25 ... Average Best Composite Predictor: Top-3.41 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 9 (0.53) Err: 5 (0.56) NErr: 0.53 WErr: 0.53 NWErr: 0.51 CScore: 0.51 Ok: 2,8,10,12 Err: 1,5,9,15,16 CompPred Top-1 - Sub: 17 Pred: 4 (0.24) Err: 4 (1.00) NErr: 0.62 WErr: 1.00 NWErr: 0.61 CScore: 0.61 Ok: Err: 6,9,13,16 CompPred Top-2 - Sub: 17 Pred: 11 (0.65) Err: 9 (0.82) NErr: 0.71 WErr: 0.78 NWErr: 0.67 CScore: 0.67 Ok: 8,12 Err: 1,3,4,6,9,11,13,15,16 CompPred Top-3 - Sub: 17 Pred: 12 (0.71) Err: 9 (0.75) NErr: 0.68 WErr: 0.72 NWErr: 0.65 CScore: 0.65 Ok: 2,8,12 Err: 1,3,4,5,9,11,15,16,17 CompPred Top-4 - Sub: 14 Pred: 10 (0.71) Err: 4 (0.40) NErr: 0.43 WErr: 0.35 NWErr: 0.41 CScore: 0.41 Ok: 2,6,7,8,10,12 Err: 1,9,16,17 CompPred Top-5 - Sub: 1 Pred: 0 (0.00) Err: 0 (�) NErr: � WErr: � NWErr: � CScore: � Ok: Err: Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.00 p(e <= 0.25) = 0.06 p(e <= 0.30) = 0.12 p(e <= 0.35) = 0.16 p(e <= 0.40) = 0.26 p(e <= 0.45) = 0.36 p(e <= 0.50) = 0.60 Computing top composite predictors at time 13.25 (without crossvalidation) ... Computing predictors at time 13.25 ... Defining composite predictors ... 1. Pred 13.0 - Pat: Pos - Sub: 17 Pred: 8 (0.47) Err: 2 (0.25) NErr: 0.38 WErr: 0.21 NWErr: 0.36 Ok: 2,5,6,8,11,12 Err: 13,16 2. Pred 15.0 - Pat: Pos - Sub: 17 Pred: 7 (0.41) Err: 1 (0.14) NErr: 0.35 WErr: 0.13 NWErr: 0.36 Ok: 3,4,10,13,15,16 Err: 9 3. Pred 1.6 - Pat: Pos - Sub: 17 Pred: 12 (0.71) Err: 5 (0.42) NErr: 0.44 WErr: 0.34 NWErr: 0.39 Ok: 5,6,7,8,10,12,14 Err: 3,4,11,16,17 4. Pred 9.0 - Pat: Pos - Sub: 17 Pred: 6 (0.35) Err: 1 (0.17) NErr: 0.38 WErr: 0.20 NWErr: 0.39 Ok: 2,6,11,12,14 Err: 1 Testing top composite predictors at time 13.25 ... Best Composite Predictor: Top-2 Score: 0.22 13.0 { RPL37A RPL38 } 15.0 { FLI1 VAMP3 } 1.6[1.3,1.5] { PHB2 FBL AKR1B1 BIN1 TRBC1 RANGRF IMPDH2 LCK } 9.0 { LOC130074 CYFIP2 } Testing top composite predictors at time 41.0 ... Average Best Composite Predictor: Top-2.88 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 13 (0.76) Err: 1 (0.08) NErr: 0.18 WErr: 0.06 NWErr: 0.15 CScore: 0.15 Ok: 1,3,5,6,7,8,10,11,12,15,16,17 Err: 2 CompPred Top-1 - Sub: 17 Pred: 6 (0.35) Err: 0 (0.00) NErr: 0.32 WErr: 0.00 NWErr: 0.30 CScore: 0.30 Ok: 1,6,7,8,10,17 Err: CompPred Top-2 - Sub: 17 Pred: 12 (0.71) Err: 0 (0.00) NErr: 0.15 WErr: 0.00 NWErr: 0.12 CScore: 0.12 Ok: 1,3,5,6,7,8,10,11,12,15,16,17 Err: CompPred Top-3 - Sub: 17 Pred: 14 (0.82) Err: 2 (0.14) NErr: 0.21 WErr: 0.12 NWErr: 0.17 CScore: 0.17 Ok: 1,3,5,6,7,8,10,11,12,15,16,17 Err: 2,13 CompPred Top-4 - Sub: 17 Pred: 15 (0.88) Err: 4 (0.27) NErr: 0.29 WErr: 0.22 NWErr: 0.25 CScore: 0.25 Ok: 1,5,6,7,8,10,11,12,15,16,17 Err: 2,9,13,14 CompPred Top-5 - Sub: 17 Pred: 15 (0.88) Err: 4 (0.27) NErr: 0.29 WErr: 0.22 NWErr: 0.25 CScore: 0.25 Ok: 1,5,6,7,8,10,11,12,15,16,17 Err: 2,9,13,14 CompPred Top-6 - Sub: 13 Pred: 12 (0.92) Err: 3 (0.25) NErr: 0.27 WErr: 0.20 NWErr: 0.28 CScore: 0.28 Ok: 1,5,6,7,8,10,12,15,17 Err: 2,3,13 Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.01 p(e <= 0.20) = 0.01 p(e <= 0.25) = 0.04 p(e <= 0.30) = 0.07 p(e <= 0.35) = 0.11 p(e <= 0.40) = 0.33 p(e <= 0.45) = 0.40 p(e <= 0.50) = 0.50 Computing top composite predictors at time 41.0 (without crossvalidation) ... Computing predictors at time 41.0 ... Defining composite predictors ... 1. Pred 3.65 - Pat: Pos - Sub: 17 Pred: 9 (0.53) Err: 0 (0.00) NErr: 0.24 WErr: 0.00 NWErr: 0.21 Ok: 1,5,6,7,8,10,11,16,17 Err: 2. Pred 15.0 - Pat: Pos - Sub: 17 Pred: 9 (0.53) Err: 0 (0.00) NErr: 0.24 WErr: 0.00 NWErr: 0.21 Ok: 1,3,6,7,8,10,12,15,17 Err: 3. Pred 8.0 - Pat: Pos - Sub: 17 Pred: 8 (0.47) Err: 0 (0.00) NErr: 0.26 WErr: 0.00 NWErr: 0.25 Ok: 1,5,6,7,11,12,14,17 Err: 4. Pred 3.61 - Pat: Pos - Sub: 17 Pred: 11 (0.65) Err: 2 (0.18) NErr: 0.29 WErr: 0.15 NWErr: 0.26 Ok: 1,5,6,7,8,12,15,16,17 Err: 2,13 5. Pred 3.31 - Pat: Pos - Sub: 17 Pred: 13 (0.76) Err: 5 (0.38) NErr: 0.41 WErr: 0.31 NWErr: 0.35 Ok: 1,5,6,7,8,10,11,12 Err: 2,9,14,16,17 6. Pred 12.0 - Pat: Pos - Sub: 17 Pred: 13 (0.76) Err: 5 (0.38) NErr: 0.41 WErr: 0.32 NWErr: 0.36 Ok: 1,5,6,7,8,12,16,17 Err: 2,3,4,11,13 Testing top composite predictors at time 41.0 ... Best Composite Predictor: Top-3 Score: 0.10 3.65[3.54] { CD46 PDE4B BACH1 } 15.0 { FLI1 VAMP3 } 8.0 { ZFYVE26 KIAA0226 } 3.61[3.60[3.59[3.57[3.56[3.55[3.52[3.50[3.49[3.44[3.43[3.42[3.41[3.40[3.39[3.35[3.34[3.33[3.32[3.30[3.29[3.28[3.27[3.26[3.25[3.24[3.17[3.16[3.15[3.14[3.12[3.9[3.7[3.5[3.3[3.2,3.0]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]]],3.58[3.53[3.51[3.48[3.47[3.46[3.45[3.38[3.37[3.36,3.23[3.22[3.21[3.20[3.19[3.18[3.13[3.11[3.10[3.6,3.8[3.1]]]]]]]]]]]]]]]]]]]] { FRY MAEA RTN3 SIRPB1 IFI44 ST6GALNAC2 TRAFD1 IFI44L TREX1 ADM CPD SIRPA DHX8 ACSL1 CD93 WDFY3 KIAA0082 KIF1B CDC2L6 TBC1D1 RAB6IP1 BICD2 KIAA0999 TDRD7 FOS PADI4 FPR2 DDX58 ALOX5AP GBP1 LAMP3 CHMP2A TOR1B PILRA HCK IFI16 IFI35 SP110 IFIT2 IFIT1 IFIT3 IL1RN IRF7 ISG20 LAMP2 CYP4F3 MMP9 MX1 MX2 ATF3 NCF2 NEDD9 NFIL3 NQO2 OAS1 OAS2 OAS3 PAK1 PAK2 ABHD5 PHF11 HERC5 PECAM1 PFKFB3 SERPINA1 PITPNA PLAUR PLSCR1 MANSC1 HERC6 TMEM140 PSEN1 C20orf3 PYGL RAF1 RALB CEACAM1 IFIH1 PARP12 SORL1 TRIM21 STX3 TLE4 TNFAIP6 C5AR1 CA4 GPR177 NPL OASL NCOA1 NUMB TNFSF10 SKAP2 MGAM RSAD2 UBE2L6 C1orf38 ISG15 CDA OSBPL2 } 3.31[3.4] { TANK BAZ1A SMCHD1 FAS NBN USP15 } 12.0 { GBP2 IRF1 } Testing top composite predictors at time 84.0 ... Average Best Composite Predictor: Top-2.06 Validation results for top composite predictors ... CompPred Best - Sub: 17 Pred: 16 (0.94) Err: 2 (0.12) NErr: 0.15 WErr: 0.11 NWErr: 0.12 CScore: 0.12 Ok: 1,3,4,5,6,7,8,9,10,11,12,14,16,17 Err: 13,15 CompPred Top-1 - Sub: 17 Pred: 9 (0.53) Err: 2 (0.22) NErr: 0.35 WErr: 0.18 NWErr: 0.32 CScore: 0.32 Ok: 5,7,8,10,11,12,17 Err: 13,15 CompPred Top-2 - Sub: 17 Pred: 15 (0.88) Err: 2 (0.13) NErr: 0.18 WErr: 0.11 NWErr: 0.16 CScore: 0.16 Ok: 1,3,4,5,7,8,9,10,11,12,14,16,17 Err: 13,15 CompPred Top-3 - Sub: 16 Pred: 15 (0.94) Err: 2 (0.13) NErr: 0.16 WErr: 0.11 NWErr: 0.15 CScore: 0.15 Ok: 1,3,4,5,6,7,8,9,10,11,12,16,17 Err: 13,15 CompPred Top-4 - Sub: 1 Pred: 1 (1.00) Err: 0 (0.00) NErr: 0.00 WErr: 0.00 NWErr: 0.46 CScore: 0.46 Ok: 6 Err: Randomization Summary ... Nullhypothesis: computed predictor can predict random classification with error rate <= x Composite Predictor Best - Tests: 100 - p(e <= 0.05) = 0.00 p(e <= 0.10) = 0.00 p(e <= 0.15) = 0.00 p(e <= 0.20) = 0.01 p(e <= 0.25) = 0.05 p(e <= 0.30) = 0.12 p(e <= 0.35) = 0.13 p(e <= 0.40) = 0.30 p(e <= 0.45) = 0.46 p(e <= 0.50) = 0.58 Computing top composite predictors at time 84.0 (without crossvalidation) ... Computing predictors at time 84.0 ... Defining composite predictors ... 1. Pred 8.0 - Pat: Pos - Sub: 17 Pred: 13 (0.76) Err: 2 (0.15) NErr: 0.24 WErr: 0.12 NWErr: 0.20 Ok: 1,5,6,7,8,9,10,11,12,14,17 Err: 13,15 2. Pred 1.13 - Pat: Neg - Sub: 17 Pred: 13 (0.76) Err: 2 (0.15) NErr: 0.24 WErr: 0.13 NWErr: 0.21 Ok: 3,4,5,6,7,8,10,11,12,16,17 Err: 13,15 3. Pred 3.1 - Pat: Pos - Sub: 17 Pred: 14 (0.82) Err: 4 (0.29) NErr: 0.32 WErr: 0.24 NWErr: 0.27 Ok: 1,4,5,6,7,8,10,12,16,17 Err: 2,13,14,15 Testing top composite predictors at time 84.0 ... Best Composite Predictor: Top-2 Score: 0.12 8.0 { ZFYVE26 KIAA0226 } 1.13[1.11[1.10[1.9[1.7[1.6[1.3,1.5]]]]]] { PHB2 FBL AKR1B1 23157-at LAT BIN1 TRBC1 RANGRF IMPDH2 LCK MAL CUTA TMEM204 SKAP1 PABPC4 CD2 FAIM3 CD27 } 3.1 { OAS3 HERC5 OASL RSAD2 UBE2L6 }