Combining DNA and protein alignments to improve genome annotation with LiftOn

Table 3.

The percentage of protein-coding transcripts mapped with varying degrees of protein sequence identity by LiftOn, Liftoff, and miniprot

Experiment LiftOn (%)
(1/≥0.8/≥0.5)
Liftoff (%)
(1/≥0.8/≥0.5)
miniprot (%)
(1/≥0.8/≥0.5)
GRCh38.p14 to CHM13 64.4/84.3/84.6 64.3/84.2/84.3 49.7/82.5/83.4
Human to Pan troglodytes 17.8/83.5/84.6 17.8/82.5/83.2 13.4/80.7/82.0
Mus musculus to Rattus norvegicus 3.6/76.8/83.5 3.4/71.8/77.0 2.7/72.4/77.4
Drosophila melanogaster to Drosophila erecta 15.2/81.7/92.0 13.2/80.1/90.3 10.2/70.0/75.4
  • Scores are categorized as 100% identity (score = 1), at least 80% identity (score ≥ 0.8), and at least 50% identity (score ≥ 0.5). The comparison evaluates the performance of the programs in mapping annotation from GRCh38.p14 to CHM13, human to P. troglodytes, M. musculus to R. norvegicus, and D. melanogaster to D. erecta.

This Article

  1. Genome Res. 35: 311-325

Preprint Server