Skip to main content
. 2018 Aug 16;46(21):e125. doi: 10.1093/nar/gky724

Table 1.

Error correction with multiple short read coverages

CH17-157L1 BAC clone – number of aligned reads
10× 20× 30×
Tool Mapped 80–90% 90–95% >95% Mapped 80–90% 90–95% >95% Mapped 80–90% 90–95% >95%
Uncorrected 33 842 17 582 1 974 461 33 842 17 582 1 974 461 33 842 17 582 1 974 461
Colormap 34 215 15 622 4 979 1 579 34 515 14 882 5 989 2 585 34 710 14 428 6 528 3 263
HALC (w/ LoRDEC) 33 933 17 656 1 870 569 35 287 13 250 5 708 5 390 37 217 12 040 6 948 8 845
HALC (w/o LoRDEC) 33 908 17 662 1 867 537 34 242 17 919 1 627 816 34 908 18 360 1 986 1 360
Hercules 36 585 14 539 6 675 3 917 40 863 14 165 7 067 8 498 41 549 13 880 7 162 9 631
LoRDEC 36 604 13 752 7 050 6 270 37 098 11 249 7 302 10 953 36 965 10 597 7 157 12 118
LSC 35 991 14 465 6 234 3 939 40 368 13 939 6 901 8 571 41 254 13 774 6 957 9 788
proovread 34 191 16 881 3 301 1 013 34 361 16 534 3 851 1 380 34 604 16 324 4 223 1 821
CH17-227A2 BAC clone - Number of aligned reads
Uncorrected 45 625 25 356 7 385 219 45 625 25 356 7 385 219 45 625 25 356 7 385 219
Colormap 45 679 22 878 9 295 1 223 45 712 21 701 9 897 2 072 45 758 21 019 10 269 2 590
HALC (w/ LoRDEC) 45 772 25 460 7 058 627 46 929 14 570 8 843 14 061 48 682 13 118 9 207 18 122
HALC (w/o LoRDEC) 45 679 22 878 9 295 1 223 45 943 26 399 6 414 656 47 038 25 223 8 453 1 833
Hercules 50 090 15 883 11 007 13 256 54 551 14 800 10 546 19 741 54 666 14 421 10 329 20 645
LoRDEC 46 467 14 787 10 010 13 053 46 893 11 422 8 301 20 332 46 956 10 329 7 719 22 597
LSC 50 162 16 206 11 727 12 385 53 762 15 071 10 805 18 568 53 887 14 624 10 578 19 501
proovread 45 700 23 444 8 869 1 139 45 797 22 612 9 289 1 833 45 919 22 044 9 467 2 489

We applied Hercules to correct PacBio reads generated from two BAC clones (CH17-157L1 and CH17-227A2) using Illumina reads from the same resource. We subsampled the Illumina reads into 10×, 20× and 30× to test the performance of the tools for both low and moderate short read coverage. We report the accuracy as the alignment identity as calculated by the BLASR (34) aligner. Mapped refers to the number of any reads aligned to the Sanger-assembled reference for these clones with any identity, where the other columns show the number of alignments within respective identity brackets.