Erroneous attribution of relevant transcription factor binding sites despite successful prediction of cis-regulatory modules

Table 2 CRM activity and conservation

CRM^a	activity	AVID	AVID+ DIALIGN	phastCons (average)	peakPhastCons (100 bp)	peakPhastCons (200 bp)	peakPhastCons (500 bp)	TFBS presence^b	TFBS mismatches^c
cooc164	+	0.6002	0.7710	0.5532	0.6651	0.6337	0.6431	0.8	0.21
cooc102	+	0.6345	0.8230	0.4246	0.4836	0.5038	0.5325	0.8	0.15
DME2	+	0.6694	0.7860	0.4472	0.5416	0.5296	0.5133	0.8	0.19
cooc310	+	0.6766	0.5250	0.6649	0.7616	0.7459	0.8099	0.8	0.26
cooc404	+	0.6874	0.7530	0.6466	0.6718	0.6740	0.6778	0.8	0.29
cooc110	+	0.6867	0.7150	0.5072	0.5808	0.5767	0.5930	0.8	0.34
	+ Average	0.6591	0.7288	0.5406	0.6174	0.6106	0.6283	0.8	0.24
	+ StdDev	0.0348	0.1061	0.1001	0.1010	0.0916	0.1089	0	0.07
DME31	-	0.5037	0.5010	0.4509	0.6249	0.4786	0.4628	0.6	0.38
DME30	-	0.5597	0.5840	0.5403	0.6183	0.5900	0.5292	0.4	0.35
DME7	-	0.5819	0.5350	0.5822	0.5609	0.5103	0.4809	0.6	0.20
DME3	-	0.6095	0.6920	0.5168	0.4784	0.4826	(<500 bp)	0.8	0.22
DME25	-	0.7401	0.8830	0.4480	0.5255	0.4964	0.4640	1	0.14
cooc437	-	0.6430	0.7840	0.5899	0.5970	0.5971	0.6145	0.6	0.12
DME4	-	nd	nd	0.5025	0.6021	0.5307	0.5200	nd	nd
	- Average	0.6063	0.6632	0.5187	0.5724	0.5265	0.5119	0.67	0.23
	- StdDev	0.0807	0.1502	0.0569	0.0540	0.0490	0.0575	0.21	0.11
p value (Wilcoxon one-sided)		0.08983	0.2424	0.4726	0.2226	0.0507	0.0206	0.06287	0.5909
with DME25 switched		0.008838	0.05303	0.6859	0.4178	0.1474	0.1010	0.005269	0.3194

^a Bold type indicates CRMs tested for activity in this paper; others are from [11].
^b TFBS presence was scored as number of sites present in D. pse. divided by number of sites in D. mel.
^cTFBS mismatches were scored as the fraction of unaligned nucleotides in each D. pse. vs. aligned D. mel. site averaged over all of the D. mel TFBSs for the CRM region (e.g., a fully conserved site = 0, a completely unaligned site = 1, average for the CRM if these were the only two sites = 0.5).

ISSN: 1471-2164