Table 1. The structure of synthetic dataset S2 (dataset with mixture distributions).

In positive marker group, each square is a 25(samples)*10(features) matrix in which each feature was generated by gamma (the red cells) or normal (the green cells) distribution function (generated by rgamma or rnorm in R). But in negative marker group, each square is a 25(samples)*900(features) matrix in which each feature was also generated by normal distribution function.

Class	Sub- class	Positive marker group										Negative marker group
		1	2	3	4	5	6	7	8	9	10
Class 1	1	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Mean
		7.18	0.61	1.70	0.81	2.36	7.18	0.61	1.70	0.81	2.36	0.14
		Rate	Rate	Rate	Rate	Rate	Rate	Rate	Rate	Rate	Rate	sd
		44.38	71.12	517	79.70	316	44.38	71.12	517	79.70	316	0.06
	2	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Shape	Mean
		6.98	0.51	1.80	0.91	2.46	6.98	0.51	1.80	0.91	2.46	0.14
		Rate	Rate	Rate	Rate	Rate	Rate	Rate	Rate	Rate	Rate	sd
		44.38	71.12	517	79.70	316	44.38	71.12	517	79.70	316	0.06
Class 2	3	Shape	Shape	Shape	Shape	Shape	Mean	Mean	Mean	Mean	Mean	Mean
		5.70	0.85	1.32	0.33	2.88	0.14	0.009	0.005	0.004	0.009	0.14
		Rate	Rate	Rate	Rate	Rate	sd	sd	sd	sd	sd	sd
		44.38	27.40	210	91.20	507	0.06	0.007	0.002	0.006	0.06	0.06
	4	Shape	Shape	Shape	Shape	Shape	Mean	Mean	Mean	Mean	Mean	Mean
		6.60	0.75	1.22	0.43	2.98	0.13	0.010	0.004	0.003	0.010	0.14
		Rate	Rate	Rate	Rate	Rate	sd	sd	sd	sd	sd	sd
		44.38	27.40	210	91.20	507	0.06	0.007	0.002	0.006	0.06	0.06