hicBuildMatrix ends with KeyError

[cgirardot]

Dear, I am running hicBuildMatrix 3.5.3 on a pretty big BAM pair (4 nextseq high lanes merged => 80G each) using the command line:

hicBuildMatrix --samFiles Fwd-Reads.bam Rev-Reads.bam --outFileName 1Kb-bin-matrix.h5 --outBam unsorted.bam --QCfolder /tmpdata/qc/. --restrictionCutFile rest_site_positions_DpnII_dm6.bed --minMappingQuality 10 --danglingSequence 'GATC' --threads 10 --binSize '1000' --restrictionSequence 'GATC'  2>error.txt 1>output.txt 

The offending key is in the BAM headers (on both Fwd & Rev bam inputs):

samtools view -H  Fwd-Reads.bam | grep chrUn_DS484472v1
@SQ SN:chrUn_DS484472v1 LN:2046

Error is :

INFO:hicexplorer.hicBuildMatrix:processing 361050129 lines took 96605.74 secs (3737.4 lines per second)

INFO:hicexplorer.hicBuildMatrix:49685436 (13.76%) valid pairs added to matrix

ERROR:hicexplorer.hicBuildMatrix:'chrUn_DS484472v1'Traceback (most recent call last):
  File "/g/funcgen/gbcs/public/software/conda/envs/hicexplorer-3.5.3/lib/python3.6/site-packages/hicexplorer/hicBuildMatrix.py", line 966, in process_data
    pRfPositions[mate_ref][frag_start: frag_end]))
KeyError: 'chrUn_DS484472v1'

Do you have an idea of what could be wrong ?

NB: Running the tech replicates individually worked in Galaxy with 3.4.3 version ; I had to move to the command line as the jobs were not ending on the cluster with the big merged files (due to time limit and me not knowing how long it needs)

[joachimwolff]

Please update to HiCExplorer 3.6. With 3.5 we introduced a mandatory restriction cut site file, however, there have been some bugs that are closed with version 3.6.

Galaxy with version 3.4 does not have this mandatory restriction cut site file and therefore has not the presented bug.

Best,

Joachim

[cgirardot]

ok thx.

[cgirardot]

solved with 3.6; thank you. Is there any chance to release 3.6 in galaxy soon ? feel free to close on reply :-)

[joachimwolff]

I hope I find time in Q1 this year.