Skip to main content

Advertisement

Table 2 Results for LoRDEC, proovread and Jabba

From: Jabba: hybrid error correction for long sequencing reads

  Gain (%) Accuracy (%) Error-free (%) Aligned (%) Throughput (%) N50 (bp)
E. coli - simulated short and real long reads - 4.7 Mbp
 Uncorrected reads   85.16 0 59.16   13,578
 LoRDECn 96.46 99.47 13.74 82.16 62.30 4661
 LoRDEC 98.83 99.82 79.31 88.95 63.70 7618
 proovread 99.64 99.94 89.64 99.57 58.65 5706
 Jabba 99.70 99.95 95.70 99.23 57.04 12,760
A. hydrophila—simulated short and simulated long reads: 4.8 Mbp
 Uncorrected reads   86.84 0 100   10,421
 LoRDECn 99.21 99.89 25.29 96.72 94.79 7625
 LoRDEC 99.93 99.99 86.74 99.76 95.35 9695
 proovread 99.99 99.99 96.53 99.99 95.40 9803
 Jabba 99.74 99.96 97.66 99.98 98.04 10,215
S. cerevisiae - simulated short and real long reads: 12.3 Mbp
 Uncorrected reads   83.21 1.50 27.99   3969
 LoRDECn 91.17 98.51 44.02 77.77 21.72 2869
 LoRDEC 92.08 98.67 60.82 83.12 30.43 3802
 proovread
 Jabba 99.87 99.97 98.35 99.93 27.67 8373
O. tauri - real short and real long reads: 13.2 Mbp
 Uncorrected reads   83.83 0.05 23.10   7322
 LoRDECn 91.04 98.55 63.60 85.05 31.43 985
 LoRDEC 91.51 98.62 66.76 85.42 31.54 1043
 proovread 98.11 99.69 80.28 90.55 26.31 1501
 Jabba 99.06 99.84 83.33 93.31 13.81 4183
A. thaliana - simulated short and real long reads: 121 Mbp
 Uncorrected reads   83.32 8.00 47.82   14,256
 LoRDEC 90.43 98.40 59.35 50.69 46.09 904
 proovread 91.11 98.51 69.71 96.66 42.08 7788
 Jabba 99.47 99.91 96.67 99.85 39.87 12,647
D. melanogaster—simulated short and real long reads: 122 Mbp
 Uncorrected reads   85.70 22.97 41.72   12,478
 LoRDEC 89.18 98.45 54.29 49.24 44.78 1119
 proovread 97.07 99.58 67.72 98.36 43.49 11,476
 Jabba 99.51 99.93 96.24 99.81 38.20 15,553
 Jabbap 99.51 99.93 96.24 99.82 38.22 15,564
  1. Results for proovread on S. cerevisiae have been left out because they did not compute in 3 days. The subscript p indicates that the tool used the reference genome instead of short reads. The subscript n indicates that the tool used uncorrected short reads