Skip to main content
. 2015 Jun 10;31(12):i80–i88. doi: 10.1093/bioinformatics/btv262

Table 2.

The performance comparison of our method on the F.tularensis dataset

Correction method Assembler MA TPR local MA TPR FPR
misSEQuel (paired-end data only) Velvet 100% (11/11) 100% (36/36) 58% (180/312)
SOAPdenovo 100% (10/10) 100% (35/35) 63% (165/263)
ABySS 100% (64/64) 100% (32/32) 87% (20/23)
SPAdes (−rr) 100% (11/11) 100% (30/30) 83% (52/63)
SPAdes (++rr) 100% (23/23) 100% (31/31) 86% (49/57)
IDBA 100% (10/10) 100% (31/31) 38% (57/149)
misSEQuel (optical mapping data only) Velvet 55% (6/11) 69% (25/36) 24% (76/312)
SOAPdenovo 80% (8/10) 63% (22/35) 29% (77/263)
ABySS 69% (44/64) 88% (28/32) 13% (3/23)
SPAdes (−rr) 91% (10/11) 87% (26/30) 21% (13/63)
SPAdes (++rr) 87% (20/23) 81% (25/31) 16% (9/57)
IDBA 90% (9/10) 77% (24/31) 10% (15/149)
misSEQuel (paired-end and optical mapping data) Velvet 55% (6/11) 100% (36/36) 22% (68/312)
SOAPdenovo 80% (8/10) 84% (21/35) 20% (53/263)
ABySS 69% (44/64) 88% (28/32) 13% (3/23)
SPAdes (−rr) 91% (10/11) 87% (26/30) 19% (12/63)
SPAdes (++rr) 97% (20/23) 81% (25/31) 16% (9/57)
IDBA 90% (9/10) 77% (24/31) 9% (14/149)
REAPR Velvet 55% (6/11) 11% (4/36) < 1% (2/312)
SOAPdenovo 20% (2/10) 14% (5/35) 2% (6/263)
ABySS 13% (8/64) 13% (4/32) 4% (1/23)
SPAdes (−rr) 27% (3/11) 27% (8/30) 5% (3/63)
SPAdes (++rr) 0% (0/23) 19% (6/31) 11% (6/57)
IDBA 40% (4/10) 13% (4/31) 4% (6/149)
Pilon Velvet 27% (3/11) 3% (1/36) < 1% (3/312)
SOAPdenovo 10% (1/10) 9% (3/35) 2% (5/263)
ABySS 3% (2/64) 6% (2/32) 4% (3/23)
SPAdes (−rr) 0% (0/11) 3% (1/30) 5% (5/63)
SPAdes (++rr) 0% (0/23) 10% (3/31) 12% (7/57)
IDBA 0% (0/10) 10% (3/31) 4% (5/149)

The TPR in this context is a contig that is misassembled and is predicted to be so. The FPR is a correctly assembled contig that was predicted to be misassembled. The TPR and FPR are given as percentages with the raw values given in brackets. Bold values emphasize the benefit of using both data sources.