Call Set Refinement (CSR) filtering step gets unresponsive for chr3:90000001-120000001 chunk

[redekarnr]

I am running into issue with Call Set Refinement (CSR) filtering in Octopus (0.7.4) (HEAD 17a597d).

Command:

octopus --threads 28 -C cancer --working-directory /gpfs/gsfs12/users/NHLBI_IDSS/projects/NHLBI-1053_207/genome-seek-out --temp-directory-prefix octopus/somatic/chunks/chr3:90000001-120000001/Sample_15_FIG430577_0001_tmp -R /data/OpenOmics/references/genome-seek/Homo_sapiens_assembly38.fasta -I /gpfs/gsfs12/users/NHLBI_IDSS/projects/NHLBI-1053_207/genome-seek-out/BAM/Sample_34_FIG470497_0001_resent.recal.bam /gpfs/gsfs12/users/NHLBI_IDSS/projects/NHLBI-1053_207/genome-seek-out/BAM/Sample_15_FIG430577_0001.recal.bam --normal-sample Sample_34_FIG470497_0001_resent -o /gpfs/gsfs12/users/NHLBI_IDSS/projects/NHLBI-1053_207/genome-seek-out/octopus/somatic/chunks/chr3:90000001-120000001/Sample_15_FIG430577_0001.vcf.gz --forest-model /data/OpenOmics/references/genome-seek/Octopus/germline.v0.7.4.forest --somatic-forest-model /data/OpenOmics/references/genome-seek/Octopus/somatic.v0.7.4.forest --annotations AC AD DP -T chr3:90000001-120000001

Slurm Log: (full logfile attached)

slurm-65136079_octosomatic.txt

` ........

[2025-08-18 10:49:05] chr3:117037291 86.4% 37m 4s 5m 48s
[2025-08-18 10:49:31] chr3:116929889 87.4% 37m 29s 5m 22s
[2025-08-18 10:49:50] chr3:118341061 88.4% 37m 48s 4m 55s
[2025-08-18 10:50:16] chr3:118060171 89.4% 38m 15s 4m 30s
[2025-08-18 10:50:38] chr3:118379144 90.4% 38m 36s 4m 4s
[2025-08-18 10:50:59] chr3:116784189 91.4% 38m 58s 3m 37s
[2025-08-18 10:51:25] chr3:118431880 92.4% 39m 23s 3m 12s
[2025-08-18 10:51:47] chr3:119372281 93.4% 39m 46s 2m 46s
[2025-08-18 10:52:10] chr3:119648935 94.4% 40m 9s 2m 20s
[2025-08-18 10:52:36] chr3:118631861 95.4% 40m 34s 1m 55s
[2025-08-18 10:53:12] chr3:90842068 96.4% 41m 11s 1m 29s
[2025-08-18 10:53:47] chr3:119620529 97.5% 41m 45s 1m 4s
[2025-08-18 10:54:42] chr3:119741167 98.5% 42m 40s 39s
[2025-08-18 11:20:12] chr3:91219060 99.5% 1h 8m 20s
[2025-08-18 11:46:33] - 100% 1h 34m -
[2025-08-18 11:46:36] Starting Call Set Refinement (CSR) filtering
[2025-08-18 11:46:36] CSR: Starting registration pass
[2025-08-18 11:46:36] ------------------------------------------------------------------------
[2025-08-18 11:46:36] current | | time | estimated
[2025-08-18 11:46:36] position | completed | taken | ttc
[2025-08-18 11:46:36] ------------------------------------------------------------------------
[2025-08-18 11:46:46] chr3:90893581 3.0% 9s 5m 13s
[2025-08-18 11:46:58] chr3:93371512 11.2% 21s 2m 50s
slurmstepd: error: *** JOB 65136079 ON cn2490 CANCELLED AT 2025-08-22T10:10:49 DUE TO TIME LIMIT ***
`

Note the difference in time stamps on the last step [2025-08-18 11:46:58] and job cancellation time [2025-08-22T10:10:49]. Processing just halts after chr3:93371512 position and job hits walltime. The position where it halted overlaps with highly repetitive centromere region of chr3 on GRCh38.

Surprisingly, this issue is only observed in 2 out 40 WGS libraries in my cohort. The CSR filtering step ran just fine for chr3:90000001-120000001 chunk in 38 other WGS libraries. I am struggling to understand what is causing this issue, and how can i adjust filtering parameters to prevent it.

Any feedback or recommendations to troubleshoot this issue would be highly appreciated. Thank you.