find-motifs fails to load before_file

[Ge0rges]

Hi @ArtRand,

On the latest (0.3.1) and previous version of modkit, the attached bedfile produces a crash in find-motifs when I run this command:

modkit find-motifs -i barcode11.bed -r contigs.fna -o  dir

Out:

methylation_5/Cognaticolwellia_r-contigs/barcode11-motifs.tsv --threads 20
> loading references from "mags/Cognaticolwellia_r-contigs.fna"
> loaded 40 sequence(s)
> Error! failed to parse any bedmethyl records

[ArtRand]

Hello @Ge0rges,

It looks like all of the bedmethyl records are being filtered out due to low coverage. I'll make the logging more descriptive in the next release.

The default value to --min-coverage is 5, and the max valid coverage in this pileup is 4. When I look at

awk '($5>=4)' barcode11.bed | less -S

You can see that most of the records have very low methylation, which is confirmed by the log:

> loaded 221 low-frequency, 51 middle-frequency and 3 high-frequency contexts, 3 modification codes, discarded 0 contexts.

If I run the following command, you get 78 motifs, but most are very ambiguous.

modkit find-motifs \
  -i ./barcode11.bed \
  -r ./Cognaticolwellia_r-contigs.fna \
  -t 10 \
  --min-sites 2 \
  --min-frac-mod 0.5 \
  --log-filepath find_motifs.log \
  --min-coverage 4

Higher coverage will certainly help.

[Ge0rges]

Got it that makes sense. So there isn't an actual error here, I'll add an exception in my script which will ignore the return value. Thanks!