Model selection and efficiency testing for normalization of cDNA microarray data

Table 2 Number of significant spot neighborhoods across the spatial layout of the arrays analyzed

Microarray	No normalization	Linear normalization	Global lowess	P-lowess	S-lowess	LIN	OLIN	OSLIN
SW 1	1,500	1,483	1,625	214	220	106	0	0
SW 2	808	831	1,068	218	113	67	0	0
SW 3	874	815	723	126	96	78	1	0
SW 4	741	757	846	76	43	49	0	0
Apo 1	1,173	1,196	1,276	913	491	755	100	1
Apo 2	521	518	801	176	74	221	3	0
Apo 3	562	576	706	334	79	258	0	0
Apo 4	770	771	1,058	177	14	176	2	0
Apo 5	670	648	844	357	222	381	5	0
Apo 6	432	432	1,003	129	106	296	10	0
Apo 7	516	526	1,258	186	88	194	17	0
Apo 8	850	833	1,202	684	342	458	61	9
Apo 9	1,596	1,621	1,780	1,105	644	798	21	5
Apo 10	707	711	896	261	108	279	3	0
Apo 11	504	484	1,306	166	87	258	11	0
Apo 12	1,313	1,323	1,144	425	288	370	12	17
Apo 13	1,357	1,368	1,155	653	394	568	41	0
Apo 14	862	1,005	987	273	108	272	82	0
Apo 15	733	743	1,004	588	241	502	97	0
Apo 16	942	985	1,347	786	333	470	47	0

The level of significance was defined by FDR = 0.01. The spot neighborhood was defined by a window of 5 × 5 spots.

ISSN: 1474-760X