Skip to main content

Table 8 Lerna results on real PacBio E.Coli K-12 reads

From: Lerna: transformer architectures for configuring error correction tools for short- and long-read genome sequencing

k Test PPL NG50
15 240.57 101,440
17 240.53 133,690
19 240.59 181,898
21 240.62 166,229
23 240.65 92,537
25 240.67 92,859
27 240.69 74,776
31 240.70 58,332
37 240.64 32,160
  1. Simulated annealing finds \(k=17\) as the best k value, which is also evident from the fact that it generated the minimum test perplexity. This value is quite close to \(k=19\) that generates the highest NG50 on assembly. Both of these k-values are highlighted in the Table.