Survival Analysis by Penalized Regression and Matrix Factorization

Table 1

21 important genes selected by lasso regression model.

UNIQID in microarray	Name	Biological function	Correlated cancer or carcinogenesis	coef. in lasso	coef. in elastic net
24432		Unknown		0.1987	0.0864
17316	RPS21	Ribosomal protein	HCC	0.1367	0.0705
15841	MYC	Transcription factor	Many cancers (e.g. DLBCL)	0.0818	0.0713
29250	AARS	tRNA synthase		0.1979	0.0874
30040	PHB2	Mitochondrial morphology	Breast cancer	0.0074	0.0356
30347	SIT1	Lymphoid cell marker		−0.0440	−0.0248
19373	HLA-DQA1	MHC class II alpha chain	DLBCL	−0.1279	−0.1115
28197	HLA-DPA1	MHC class II alpha chain	DLBCL	−0.0920	−0.0799
24396	HLA-DRB1	MHC class II beta chain	DLBCL	−0.1345	−0.0877
31957	CD22	B-cell receptor signalling	DLBCL, cancer drug	−0.0895	−0.0597
27091	ST6GAL1	Glycosyltransferase	Colorectal cancer	0.1062	0.0741
31316	FCRL3	New CD molecule		0.0324	0.0207
27379	LRMP	Germinal center marker		−0.0562	−0.0499
26361		Unknown		−0.0070	−0.0037
17723	IGKC	Immunoglobulin light chain		0.0341	0.0360
34407	PTPN6	Protein tyrosine phosphatase	Anaplastic large-cell lymphoma	0.0611	0.0434
24400	MGLL	Monoglyceride lipase		0.0131	0.0216
24395	IFI30	MHC class II Ag processing		−0.0889	−0.0635
16972	TXNIP	Interact with thioredoxin	Tumor suppressor gene	0.0659	0.0372
34814	IL23A	Cytokine	Oncogene or tumor suppressor gene	−0.0114	−0.0128
17475	HSPA1A	Heat shock protein	Many cancers	−0.0211	−0.0166