Table 11 Comparison of assembly results with and without error correction of S. aureus reads, including the NB model performance
From: MAC-ErrorReads: machine learning-assisted classifier for filtering erroneous NGS reads
Dataset | S. aureus reads | S. aureus reads corrected by lighter | NB correct classified reads | NB correct classified reads + lighter for correcting erroneous classified reads |
|---|---|---|---|---|
Metrics | ||||
# contigs | 3075 | 3086 | 2911 | 3084 |
Largest contig | 22,555 | 22,555 | 20,317 | 22,555 |
N50 | 2952 | 2938 | 2373 | 2940 |
NG50 | 5832 | 5867 | 4329 | 5867 |
Genome fraction (%) | 92.170 | 92.515 | 94.206 | 92.490 |
#misassemblies | 1 | 1 | 0 | 1 |
#misassembled contigs | 1 | 1 | 0 | 1 |
Misassembled contigs length | 1826 | 1826 | 0 | 1826 |
# local misassemblies | 3 | 3 | 2 | 3 |
#mismatches per 100 kbp | 24.42 | 25.26 | 21.26 | 24.49 |
Total aligned length | 2,682,193 | 2,691,757 | 2,741,617 | 2,690,983 |
#mismatches | 655 | 680 | 583 | 659 |
#indels | 17 | 19 | 10 | 19 |