Skip to main content

Table 5 Mean AUROC on test sets from cross-validation for the ablation study

From: A fair experimental comparison of neural network architectures for latent representations of multi-omics for drug response prediction

Drug

Omics stacking

Complete integration

Without integration

Integration

Without triplet loss

Gemcitabine TCGA

0.630 ± 0.055

0.605 ± 0.076

0.646 ± 0.045

0.601 ± 0.084

Gemcitabine PDX

0.640 ± 0.089

0.656 ± 0.062

0.651 ± 0.071

0.593 ± 0.044

Cisplatin

0.742 ± 0.050

0.757 ± 0.061

0.722 ± 0.066

0.734 ± 0.091

Docetaxel

0.775 ± 0.089

0.759 ± 0.031

0.772 ± 0.077

0.813 ± 0.024

Erlotinib

0.696 ± 0.101

0.744 ± 0.098

0.754 ± 0.114

0.662 ± 0.112

Cetuximab

0.748 ± 0.048

0.679 ± 0.052

0.731 ± 0.090

0.721 ± 0.055

Paclitaxel

0.695 ± 0.104

0.634 ± 0.114

0.667 ± 0.138

0.522 ± 0.085

Rank

2.00

2.57

2.14

3.29

  1. Best results are shown in bold and second best are italic. The values represent the means and standard deviations over five iterations