ERROR: Requested column 22, but database file - only has fields 1 - 0

[suyanan]

subtask.sh.e.log

use docker command:

docker run --user $(id -u):$(id -g) --rm \
  -v "${indir}":"/fiberseq" \
  -v "${reference_dir}":"/reference" \
  ${FIRE_image} \
  /bin/bash -c "pixi run fire --configfile /fiberseq/run_FIRE/config.yaml

get results:

only 4 peaks results.

I guess, there no some chroms's temp/F-2/fire-peaks/v0.1-chr*-fire-peaks.bed.gz files:

I look forward to your reply.

[cademirch]

Could you share:

1. Your config.yaml (especially the chroms or keep_chromosomes section if using)
2. Your reference genome index
3. The output of tabix -l on the pileup BED

That should help me understand what is going on here.

[suyanan]

config:

$cat config.yaml 
ref: /reference/arahy.Tifrunner.gnm2.J5K5.fa
ref_name: arahy.Tifrunner.gnm2.J5K5
manifest: /fiberseq/run_FIRE/config.tbl
$ cat config.tbl
sample	bam
h3	/fiberseq/run_FIRE/hiphase_bam/h3.hiphase.bam

reference is GCF_003086295.3_arahy.Tifrunner.gnm2.J5K5 from NCBI, and use samtools faidx to generate its fai file.

I only get 4 peaks result refering to tabix.peaks_result.txt file, and pipeup refering to tabix.pileup_result.txt.

tabix.peaks_result.txt
tabix.pileup_result.txt

[cademirch]

I wonder if this is related to you previous issue #52 which I pushed a fix for. Could you pull the latest from main and rerun the workflow?

[suyanan]

I use latest to test again, same ERROR occurs.

Also, I send you one email to your gmail, include this sample test data.

[cademirch]

Hi @suyanan I responded to your email. Let's keep the discussion here moving forward please. In any case, looking at your pileup tabix output, I see that you have many unplaced scaffolds (e.g NW_027255...), which is also the problematic contig in your above screenshots.

The fix is to restrict the chromosome set via keep_chromosomes in your config.yaml, as documented here to exclude these scaffolds from the analysis. Something like:

keep_chromosomes: "chr[0-9]+$"

would skip the scaffolds and only run on chr1–chr20.

Alternatively you can use min_contig_length to filter by size.

After updating, please clear the stale outputs and rerun:

rm -rf temp/h3 results/h3
pixi run fire --configfile config.yaml

[suyanan]

hi,
thank a lot.

Is this limitation (keep_chromosomes and min_contig_length) mandatory?
peak calling only meaningful when they are covered on main chromosomes? Should all scaffolds and contigs be removed later?

Use email to supply you test data (file size is limited) for reproducing and debugging the problem.

About your response while 4 peaks is smaller than I would expect, I would caution using the number of peaks in a human sample as a point of reference. in email, my human sample is 32.52X and peanut sample is 32.23X. Peanut genome is 2.6G.

Besides, I test tomato 13.15X depth with SLM_r2.1 version genome and broomcorn 13.67 depth with Sorghum_bicolor_NCBIv3 version genome. Both get 41 peaks and 14 peaks respectively and same error.

I test peanut sample again, use config.yaml below, also get same error and same peaks number (4 peaks):

ref: /reference/arahy.Tifrunner.gnm2.J5K5.fa
ref_name: arahy.Tifrunner.gnm2.J5K5
min_contig_length: 0
keep_chromosomes: chr[0-9XY]+$
manifest: /fiberseq/run_FIRE/config.tbl

Recommend that you test by my test data from email.

Will this error affect the peak calling results? only 4 peaks, normal?

Thanks.

[cademirch]

Is this limitation (keep_chromosomes and min_contig_length) mandatory?

No it is not mandatory, but unless you have specific reasons for including these scaffolds, I would strongly suggest excluding them. Especially since it seems like they are causing errors in the workflow currently.

Let's get the workflow to run all the way without error then we can assess the number of peaks.

[suyanan]

hi,

I test peanut sample again, also get same error.
how to adjust paras or need to fix?

[suyanan]

hi,
Has the issue been fixed?