Varia: a tool for prediction, analysis and visualisation of variable genes

BMC Bioinformatics

Table 2 Analysis of database hits and gene predictions from Varia_VIP analysis of DBLα-ζ tags. Sample tags from 971 different DBLα-ζ domains were extracted from 15 P. falciparum genomes

Domain	No. tags tested	Hit rate (%)			Average No. of clusters			Percentage correctly annotated genes (any cluster)			Percentage correctly annotated genes (top 5 Clusters)			Percentage perfect DNA sequence hits (top 5 clusters)
Seq. ID threshold →	No. tags tested	99%	95%	90%	99%	95%	90%	99%	95%	90%	99%	95%	90%	99%	95%	90%
DBLα	293	95	99	100	14	34	92	72	75	78	66	62	53	29	27	22
DBLβ	127	94	96	97	13	41	100	73	73	73	69	65	53	39	35	27
DBLδ	256	96	99	100	15	34	60	74	74	77	69	63	55	33	29	26
DBLγ	138	91	93	93	18	61	146	71	70	71	64	55	44	41	36	24
DBLε	109	78	78	78	21	61	120	54	56	56	51	44	39	25	17	13
DBLζ	48	96	100	100	29	99	263	60	63	63	54	38	29	35	21	10

Tags were run through Varia_VIP using a length filter of 150 base pairs and an identity filter of 99%, 95% and 90%. The hit rate shows the proportion of queried tags that had one or more hits in the var gene database. The average number of clusters into which hit genes were grouped and the proportion of genes for which a correct domain subtype annotation was found in any cluster, or in the five clusters with most hit sequences (top 5), is shown. Also shown is the proportion of genes for which a sequence matching the reference gene by 99% identity over at least 80% of the full sequence was found among the top five Varia_VIP clusters

ISSN: 1471-2105