Priority-Lasso: a simple hierarchical approach to the prediction of clinical outcome using multi-omics data

BMC Bioinformatics

Table 2 Variables selected by priority-Lasso in scenarios pl2A and pl2B

Block	Variable	Coef. pl2A	Coef. pl2B
1	t(8;21)(q22;q22)	-1.0289	-1.0289
	inv(16)(p13.1q22)	-1.5444	-1.5444
	NPM1 mut/FLT3-ITD neg or low	-1.0181	-1.0181
	biCEBPA	-1.2240	-1.2240
	NPM1 wt/FLT3-ITD pos or low	-0.4358	-0.4358
	t(9;11)(p21;q23)	0.4635	0.4635
	Other aberrations	-0.4376	-0.4376
	KMT2A rearrangements	-0.5440	-0.5440
	Complex karyotype	0.2970	0.2970
	Monosomal karyotype	0.0313	0.0313
	NPM1 wt/FLT3-ITD pos	0.1712	0.1712
	RUNX1 mutations	0.3065	0.3065
	ASXL mutations	-0.1224	-0.1224
	TP53 mutations	0.4306	0.4306
2	Age	0.2957	0.2617
	Sex	-0.1011
	ECOG (> 1)	0.3147	0.3206
	WBC	0.0990	0.0589
	LDH	0.1681	0.2371
	Hb	0.0700	0.0671
	PLT	-0.0960	-0.0578
4	ZBTB37	0.0047	0.0025
	MFI2	0.0090
	SH3PXD2B	0.0013	0.0418
	PDK3	-0.0187
	FAM24B	0.0248
	SIK3	-0.0063
	OR7A17	0.0039
	TBC1D17	-0.0172
	PHGDH		0.0488
	FAM171B		0.0134
	FGD5		0.0359
	F12		0.0238
	IRX1		-0.0090
	FAM92A1		0.0239
	DDIT4		0.0769
	HSPA2		0.0169

Column 1: priority of the block the variable is included in. Column 2: variable name. Column 3 and 4: coefficient of the variable in the Cox Lasso model. Variables from the block of priority 4 also appearing in Table 1 are marked in bold

ISSN: 1471-2105