Table 3.

Datasets Summary

Data set	Number of the case and control samples	Number of the features (after selection)	Reference
GSE14520	225, 220	135	Roessler S, Jia HL, Budhu A, Forgues M et al. A unique metastasis gene signature enables prediction of tumor relapse in early-stage hepatocellular carcinoma patients. Cancer Res 2010 Dec 15;70 [24]:10202-12. PMID: 21,159,642
GSE25097	268, 243	135	Tung EK, Mak CK, Fatima S, Lo RC et al. Clinicopathological and prognostic significance of serum and tissue Dickkopf-1 levels in human hepatocellular carcinoma. Liver Int 2011 Nov;31 [10]:1494 − 504. PMID: 21,955,977
GSE36376	240, 193	135	Lim HY, Sohn I, Deng S, Lee J et al. Prediction of disease-free survival in hepatocellular carcinoma by gene expression profiling. Ann Surg Oncol 2013 Nov;20 [12]:3747-53. PMID: 23,800,896
Cambridge Baby Growth Study (CBGS) cohorts	CBGS-1 85 HM, 87 FM; CBGS-2 43 HM, 25 FM, 27 HM + FM; POPS 16 HM, 11 FM	218	Acharjee, A., Prentice, P., Acerini, C. et al. The translation of lipid profiles to nutritional biomarkers in the study of infant metabolism. Metabolomics 13, 25 (2017). 10.1007/s11306-017-1166-2

Summary of public datasets used in this study. Reference refers to the earliest citation present on the corresponding NCBI Gene Expression Omnibus [28] page. Number of features reflects the input data to the corresponding GAN model