Improved ChIP-chip analysis by a mixture model approach

BMC Bioinformatics

Table 2 Comparison of different methods for the simulated data set with 2,100/2,058 spike-in regions for unamplified/amplified samples, respectively

	Mixer (Lowess)		Mixer (MA2C)		MA2C		TileMap (p = 0.01, q = 0.05)		TileMap (p = 0.03, q = 0.08)		HGMM
Sample	D	FDR	D	FDR	D	FDR	D	FDR	D	FDR	D	FDR
GSM254930	2159	0.23	1694	0.17	2219	0.28	1475	0.004	1619	0.004	1605	0.03
GSM254971	1965	0.21	2033	0.22	2187	0.30	1395	0.003	1578	0.006	1577	0.03
GSM254972	2015	0.19	2226	0.27	2151	0.27	1553	0.003	1713	0.009	1553	0.03
GSM254973	1575	0.14	1929	0.19	2094	0.28	1334	0.004	1504	0.008	1520	0.02
GSM254805	1982	0.30	1764	0.24	1671	0.27	1034	0.013	1140	0.019	939	0.03
GSM254806	2180	0.27	2344	0.33	1910	0.23	1404	0.008	1687	0.027	1372	0.03
GSM254807	1495	0.14	1926	0.18	2034	0.27	1486	0.003	1655	0.009	1519	0.03

See main text for the simulation methods. Approximately 4.3% of the probes are from spike-in regions.

ISSN: 1471-2105