I believe that my issue is similar to #12
I have a sparse matrix file of the entire genome in hicpro format.
The resolution of my matrix is 1kb. Genome size is 379.6Mb. So I have a sparse matrix file with 379633 bins.
When I run armatus with this command:
armatus -S -g 0.9 -r 1000 -s 0.1 -i az.hicpro -o az_armatus
I get an output with 72 consensus domains, which is too low (I found around 4000 TADs with HiCExplorer). The log file:
Reading input from az.hicpro.
Building matrix for chromosome N/A at resolution 1000bp with 380 rows.
Initializing matrix to zero elements
0.0459572%
0.0919145%
I read the comment to #12 and thought that maybe armatus thought my bin numbers were chromosomal coordinates and divided them by resolution, hence low number of domains. I tried the command without -r:
armatus -S -g 0.9 -s 0.1 -i az.hicpro -o az_armatus1
I got an error:
Reading input from az_mrg.hicpro.
Building matrix for chromosome N/A at resolution 1bp with 379634 rows.
Initializing matrix to zero elements
terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc
/opt/slurm/data/slurmd/job27956122/slurm_script: line 14: 2093009 Aborted (core dumped) armatus -S -i az.hicpro -g 0.9 -o az_armatus
I believe that my issue is similar to #12 I have a sparse matrix file of the entire genome in hicpro format. The resolution of my matrix is 1kb. Genome size is 379.6Mb. So I have a sparse matrix file with 379633 bins.
When I run armatus with this command:
armatus -S -g 0.9 -r 1000 -s 0.1 -i az.hicpro -o az_armatus
I get an output with 72 consensus domains, which is too low (I found around 4000 TADs with HiCExplorer). The log file:I read the comment to #12 and thought that maybe armatus thought my bin numbers were chromosomal coordinates and divided them by resolution, hence low number of domains. I tried the command without -r:
armatus -S -g 0.9 -s 0.1 -i az.hicpro -o az_armatus1
I got an error:How can I resolve this?