VarDetect: a nucleotide sequence variation exploratory tool

BMC Bioinformatics

Table 1 Comparison of the efficiency of VarDetect, PolyPhred, Genalys, novoSNP and Mutation Surveyor.

Gene (contigs)	Verified SNPs	VarDetect			PolyPhred			Genalys			novoSNP			Mut. Surveyor
		TP	FP	FN	TP	FP	FN	TP	FP	FN	TP	FP	FN	TP	FP	FN
ACOX2 (5)	10	10	28	0	3	1	7	6	277	4	9	352	1	6	40	4
ADM (2)	2	1	4	1	0	0	2	1	260	1	2	220	0	1	11	1
ARRB1 (6)	16	15	7	1	9	1	7	16	30	0	15	58	1	13	13	3
CACNA1D (11)	26	23	9	3	12	4	14	20	363	6	26	361	0	22	60	4
CACNB3 (3)	6	5	4	1	4	1	2	5	191	1	5	308	1	3	51	3
CCL2 (2)	3	3	10	0	1	0	2	3	196	0	3	130	0	2	34	1
CCL3 (2)	12	11	4	1	4	0	8	9	171	3	12	123	0	9	31	3
CCL4 (2)	10	8	11	2	6	3	4	8	136	2	8	96	2	7	20	3
CCL5 (2)	3	3	2	0	3	1	0	3	28	0	3	21	0	2	3	1
CCR7 (2)	2	2	3	0	1	0	1	2	75	0	2	129	0	1	3	1
ITGAM (13)	27	19	20	8	17	7	10	19	552	8	23	618	4	22	63	5
ITGAX (15)	25	24	28	1	13	4	12	22	704	3	24	1166	1	19	131	6
ITGB7 (9)	16	15	13	1	11	1	5	15	435	1	16	521	0	11	32	5
LIPG (1)	4	2	2	2	1	2	3	3	82	1	3	102	1	3	2	1
NPY (2)	9	7	4	2	5	0	4	8	228	1	9	334	0	7	11	2
Total 15 genes	171	148	149	23	90	25	81	140	3728	31	160	4539	11	128	505	43
77 contigs
Precision (%)		49.83			78.26			3.62			3.40			20.22
Recall (%)		86.55			52.63			81.87			93.57			74.85
F-score (%)		63.25			62.94			6.93			6.56			31.84

Our testing set comprises 77 exonic contigs from 15 human genes. Of these 77 contigs, there are 171 SNPs experimentally validated by the experts [14]. novoSNP has lowest false negative result while generates heighest false positive result (4539). The number of detected SNPs and the false positive result identified by VarDetect are better than Genalys, novoSNP and Mutation Surveyor. Although VarDetect has worse performance on false positive, its SNP identification coverage (148 out of 171) is much higher than that of PolyPhred (90 out of 171) resulting in higher false positive and lower false negative result than PolyPhred' s. From these experiments, VarDetect had the best F-score (63.25%) that slightly greater than PolyPhred (62.94%). Mutation Surveyor, Genalys and novoSNP had low F-score that had 31.84%, 6.93% and 6.56%, respectively.

ISSN: 1471-2105