Skip to main content

Table 4 Rhodobacter sphaeroides on GAGE contig assemblies using the wide MP library

From: BESST - Efficient scaffolding of large fragmented assemblies

  BESST OPERA SOPRA SSPACE Unscaffolded
  CorrEsize (kbp) err CorrEsize (kbp) err CorrEsize (kbp) err CorrEsize (kbp) err CorrEsize (kbp)
ABySS 17,6 5 19,7 51 6,7 4 9,9 1 6,9
Allpaths-LG 318,1 0 314,5 1 44,4 0 70,3 2 35,9
Bambus2 460,6 0 267,1 0 93,9 0 137,9 0 16,2
CABOG 199,8 3 97,3 4 22,3 0 33,5 0 21,5
MSR-CA 192,1 1 203,7 6 24,0 0 38,1 1 21,6
SGA 4,3 0 13,3 111 3,3 12 6,5 3 3,2
SOAPdenovo 756,7 0 720,5 10 156,2 0 206,4 2 18,6
Velvet 350,2 2 62,3 4 17,8 3 30,9 3 16,7
SUM 2299,3 11 1698,4 187 368,7 19 533,6 12  
  1. The numbers in bold face style indicate the best corrected E-size and number of errors among the stand-alone scaffolders for each assembly. Note that the corrected E-size for SOPRA is slightly less than the corrected contig size in the ABySS assembly. This can happen for low contiguity scaffolded assemblies that contains more bases than the contig assemblies (5,0Mbp and 4,5Mbp respectively on this instance). The difference in number of bases is due to the facts that GAGE evaluation script only compute statistics on contigs and scaffolds that are longer than 200bp. GAGE evaluation script returned an error when computing statistics for seven, two and one scaffolds on SSPACE results of ABySS, CABOG and MSR-CA assemblies respectively. On SOPRA and OPERA results, 1 respectively 3 scaffolds of the SGA assembly returned this error. We removed these scaffolds from the evaluation in order to compute the results. In all cases, the scaffolds removed summed up to a total length of less than 110 kbp. Thus, this has a negligible (either positive or negative) effect on E-size computation and an eventual positive effect on the number of errors. Results for BESST contained no scaffolds giving this error.