Evaluation of transfer learning for enhancing model training in SyNG-BTS using microRNA-seq data (Panel A) and RNA-seq data (Panel B) with marker filtering. (A) Evaluation metrics on the congruence of the augmented data and the empirical data, including 1 – Pct(0-markers), cARI, and CCCPCC, are calculated when pilot data are drawn from the TCGA SKCM microRNA-seq study and models are pretrained using the TCGA LAML study or the combination of the TCGA BRCA, LAML, and PRAD studies (left column of sub-panels), and when pilot data are drawn from the TCGA BRCA microRNA-seq study and the models are pretrained using the TCGA PRAD or the combination of the TCGA SKCM, LAML, and PRAD studies (right column of sub-panels). (B) Evaluation metrics for data congruence, including 1 – Pct(0-markers) and cARI, are calculated when pilot data are drawn from the TCGA BRCA RNA-seq study and the models are pretrained using the TCGA PRAD study (left column of sub-panels), and when pilot data are drawn from the TCGA PRAD RNA-seq study and the models are pretrained using the TCGA BRCA study (right column of sub-panels).