balanced contact value missing in output of cooler dump

[liu-zhiyang]

Hi,
I tried to use cooler dump --join -b -H mycool.mcool::resolutions/10000 to output balanced contacts, however, the balanced column was empty, here is the first 10 lines of the output:

chrom1 start1 end1 chrom2 start2 end2 count balanced
chr1 10000 15000 chr1 201630000 201635000 1
chr1 10000 15000 chr10 17605000 17610000 1
chr1 10000 15000 chr13 71220000 71225000 1
chr1 30000 35000 chr1 68180000 68185000 1
chr1 30000 35000 chr4 99520000 99525000 1
chr1 30000 35000 chr7 71555000 71560000 1
chr1 30000 35000 chr8 119750000 119755000 1
chr1 30000 35000 chr9 36735000 36740000 1
chr1 30000 35000 chr13 73605000 73610000 1

and this is attrs of my cool file:

'@attrs':
bin-size: 10000
bin-type: fixed
creation-date: 2025-07-25 20:37:22.986581
format: HDF5::Cooler
format-url: https://github.com/open2c/cooler
format-version: 3
generated-by: cooler-0.9.3
genome-assembly: unknown
metadata: {}
nbins: 310114
nchroms: 195
nnz: 230175878
storage-mode: symmetric-upper
sum: 658912601
bins:
'@attrs': {}
chrom:
'@attrs': {}
end:
'@attrs': {}
start:
'@attrs': {}
weight:
'@attrs':
cis_only: false
converged: true
divisive_weights: false
ignore_diags: 2
mad_max: 5
min_count: 0
min_nnz: 10
scale: 4220.152734945097
tol: 1.0e-05
var: 8.085986609393021e-06
chroms:
'@attrs': {}
length:
'@attrs': {}
name:
'@attrs': {}
indexes:
'@attrs': {}
bin1_offset:
'@attrs': {}
chrom_offset:
'@attrs': {}
pixels:
'@attrs': {}
bin1_id:
'@attrs': {}
bin2_id:
'@attrs': {}
count:
'@attrs': {}

I wonder how to solve this problem, thank you very much!

[Phlya]

How did you create this cooler? The order of pixels in the table looks very strange, like they are not sorted by chrom2, start2, end2... A missing bin on the first side is weird, and also you are requesting resolutions/10000, but actual bin size is 5000? This is all very weird.

[liu-zhiyang]

Thanks for your reply.
Sorry, I failed to paste the correct head lines of the cooler dump output, and this is the correct one:

chrom1 start1 end1 chrom2 start2 end2 count balanced
chr1 10000 20000 chr1 201630000 201640000 1
chr1 10000 20000 chr10 17600000 17610000 1
chr1 10000 20000 chr13 71220000 71230000 1
chr1 30000 40000 chr1 68180000 68190000 1
chr1 30000 40000 chr4 99520000 99530000 1
chr1 30000 40000 chr7 71550000 71560000 1
chr1 30000 40000 chr8 119750000 119760000 1
chr1 30000 40000 chr9 36730000 36740000 1
chr1 30000 40000 chr13 73600000 73610000 1

I created my cool file using the following commands:
pairtools select "(mapq1>=30) and (mapq2>=30)" mypairs.nodup.pairs.gz | cooler cload pairs -c1 2 -p1 3 -c2 4 -p2 5 ~/reference/genome/GRCh38/GRCh38.chrom.sizes:1000 - mycool.mapq_30.1000.cool
cooler zoomify --balance --balance-args '--convergence-policy store_nan' -n 16 -o mycool.mcool -r 1000,5000,10000,25000,50000,100000,250000,500000,1000000,2500000 mycool.mapq_30.1000.cool,
and due to the problem above, I re-balanced the file using for i in `cooler ls mycool.mcool`; do cooler balance $i -f --max-iters 500 -p 24; done

[Phlya]

Can you check some pixels much lower in the table? Considering that the first bin chr1 10000 20000 has only 3 total contacts, it was filtered out during balancing. Often at the ends of chromosomes the coverage is very low due to repeats etc, better to check somewhere in the middle...

[liu-zhiyang]

You are right. I checked chr1:10000000-20000000, and here is the head lines:

chrom1 start1 end1 chrom2 start2 end2 count balanced
chr1 10000000 10010000 chr1 10000000 10010000 581 0.180829
chr1 10000000 10010000 chr1 10010000 10020000 209 0.0784174
chr1 10000000 10010000 chr1 10020000 10030000 107 0.042498
chr1 10000000 10010000 chr1 10030000 10040000 60 0.0187432
chr1 10000000 10010000 chr1 10040000 10050000 29 0.0109136
chr1 10000000 10010000 chr1 10050000 10060000 25 0.00928175
chr1 10000000 10010000 chr1 10060000 10070000 18 0.00700388
chr1 10000000 10010000 chr1 10070000 10080000 25 0.00912489
chr1 10000000 10010000 chr1 10080000 10090000 37 0.0118149

Thank you very much!