Non-specific filtering of beta-distributed data

BMC Bioinformatics

Table 3 Misclassification rate of RPMM cluster analysis to find 2 groups using different variable filtering methods (top 1000 features)

	Data set 1	Data set 2	Data set 3	Data set 4	Data set 5	Data set 6	Data set 7
Tissue type	Colon cancer	Glioblastoma	Glioblastoma	Kidney	Kidney	Breast	Breast
Platform	HM27	HM27	HM450	HM27	HM450	HM27	HM450
# of samples	20 non-CIMP vs. 6 CIMP	74 non-CIMP vs. 12 CIMP	93 non-CIMP vs. 6 CIMP	50 KIRC vs. 45 non-cancer	283 KIRC vs. 160 non-cancer	37 Breast cancer vs. 20 non-cancer	56 Breast cancer vs. 17 non-cancer
No filter	0.31	0.22	NA	0	NA	0.12	NA
Filter top 1000 by:
Random *	0.34	0.27	0.40	0.004	0.005	0.12	0.20
SD-b	0.19	0.07	0.49	0	0.02	0	0.12
SD-m	0.12	0.07	0.42	0.02	0.03	0.12	0.08
MAD	0.38	0.35	0.49	0	0.005	0	0.14
DIP	0.23	0.36	0.45	0	0.005	0	0.14
Precision	0.08	0	0.10	0.03	0.01	0.11	0.22
BQ-GOF	0.19	0	0.07	0	0.01	0.25	0.23
TM-GOF	0.08	0.02	0.06	0.36	0.47	0.44	0.49
TQ-GOF	0.08	0.03	0.06	0.35	0.47	0.44	0.48
BR	0.12	0.02	0.11	0.02	0.02	0.23	0.19
AR	0.08	0.06	0.11	0.02	0.02	0.25	0.19
WAR	0.12	0.07	0.45	0.02	0.01	0.11	0.10
SD-b + TM-GOF**	0.08	0.07	0.20	0.05	0.01	0.26	0.36

ISSN: 1471-2105