search

raphael-group / hatchet

HATCHet (Holistic Allele-specific Tumor Copy-number Heterogeneity) is an algorithm that infers allele and clone-specific CNAs and WGDs jointly across multiple tumor samples from the same patient, and that leverages the relationships between clones in these samples.
BSD 3-Clause "New" or "Revised" License
68 stars 32 forks source link

Error when no reads are recorded in tumor.bed before/after centromere location #198

Open ronkesm opened 11 months ago

ronkesm commented 11 months ago

Hi,

In a number of my samples (which I am currently running through HATCHet2 individually), I am experiencing a bug that appears to occur when either pre- or post-centromeric reads on specific chromosomes aren't recorded in the baf/tumor.1bed file.

Here's an example of an error message:

Traceback (most recent call last): File "/data/home/hmz251/.conda/envs/HATCHet2/lib/python3.9/multiprocessing/pool.py", line 125, in worker result = (True, func(*args, **kwds)) File "/data/home/hmz251/.conda/envs/HATCHet2/lib/python3.9/multiprocessing/pool.py", line 48, in mapstar return list(map(*args)) File "/data/BCI-WrenchLab/globus/WGS/BAM/BASE_RECALIBRATED/hatchet-2.0.0/src/hatchet/utils/count_reads.py", line 539, in run_chromosome_wrapper run_chromosome(*param) File "/data/BCI-WrenchLab/globus/WGS/BAM/BASE_RECALIBRATED/hatchet-2.0.0/src/hatchet/utils/count_reads.py", line 535, in run_chromosome raise e File "/data/BCI-WrenchLab/globus/WGS/BAM/BASE_RECALIBRATED/hatchet-2.0.0/src/hatchet/utils/count_reads.py", line 509, in run_chromosome last_idx_p = np.argwhere(thresholds > centromere_start)[0][0] IndexError: index 0 is out of bounds for axis 0 with size 0

...Affecting chromosome 5 in this sample:

Error in chromosome chr5: index 0 is out of bounds for axis 0 with size 0

When I inspect the baf/tumour.1bed file, I see that the last read on chr5 is
chr5 34292229 PD44709a 29 35

Which is before the hg38 coordinates for the beginning of the centromere on chr5. There are no reads past the centromere. It's a strange bug but it occurs in 1/3 of the samples I run through HATCHet2 (so around 10-12).

Haven't had any issues calling CNs with any other callers on this sample (or the others where this issue appears). Sequenza output below:

image

Happy to provide a subset of the sample to reproduce.

istvankleijn commented 2 months ago

Did you manage to resolve this? I am encountering the same issue on three chromosomes in the sample I am investigating.

I also get many warnings [W::tbx_parse1] Coordinate <= 0 detected. Did you forget to use the -0 option? slightly prior to the error, but these do not seem to be specific to the three chromosomes with SNPs called only in part of one arm.

ronkesm commented 2 months ago

Hi @istvankleijn,

No, unfortunately not... this was a while ago, but I remember trying to hack together a couple of workarounds (like manually imputing post-centromeric reads into the bed file) but it ended up crashing downstream of that step anyway. It doesn't seem to be strictly related to sample quality either.

I got HATCHet2 to work in samples where I already had prior confirmation of WGD or suspected WGD (eg. from cytogenetics reports or VAF distributions), so I left it at that and ran another copy-number caller on the remaining samples I had.