Cross-species protein sequence and gene structure prediction with fine-tuned Webscipio 2.0 and Scipio

BMC Research Notes

Table 2 Test scenario 1: Reconstruction of the Loxodonta africana dynein heavy chain genes in the whole genome sequence based on human protein sequences

Tool	Predicted genes	Missing exons¹	Wrong exons²	Exon sens. %	Exon sens. (ov.)³ %	Exon spec. %	Nucl. sens. %	Nucl. spec. %	Execution time per prot. seq.
Scipio 1.5⁴	16	11	6	93.4	99.1	93.3	98.7	99.8	70 m 46s
Exonerate⁵	2145	6	5669	94.8	99.5	16.5	99.7	18.5	123 m 23s
Exonerate⁶	16	287	62	73.4	76.1	90.2	76.1	94.3	121 m 27s
Augustus⁷	1374928	0	3909434	47.9	100.0	0.0	100.0	0.3	> 10 days
BLAT⁸	-	9	264228	19.6	99.3	0.1	97.4	2.6	7 m 24s
Scipio 1.0⁴	16	14	46	86.1	98.8	83.2	97.9	99.4	8 m 24s

¹ Number of annotated exons, which are not overlapped by any predicted exon
² Number of predicted exons, which are not overlapped by any annotated exon
³ Number annotated exons, which are overlapped by at least one predicted exon divided by the number of annotated exons
⁴ Mammalia cross species default options (for detailed parameters see Additional file 5)
⁵ Parameters: --model protein2genome
⁶ Parameters: --model protein2genome --bestn 1
⁷ Parameters: --species = human --genemodel = exactlyone (for more parameters see Additional file 5)
⁸ Parameters as in ⁴: -tileSize = 7 -minIdentity = 54 -minScore = 15 -oneOff = 1

ISSN: 1756-0500