SemiBin icon indicating copy to clipboard operation
SemiBin copied to clipboard

SemiBin doesnt complete bining. Seems to be stuck

Open HitMonk opened this issue 1 year ago • 4 comments

Hello, So I have been trying to run SemiBin2 on an assembled set of contigs. however, it seems that after generating the training model there is no progress made. I left the program run for 2 days and still do not see any new files being created. The processors also dont seem to be doing anything. I also ran it again with the generated model, and the same thing happens. The one thing that i can think might be an issue is memory. As the program starts the binning part, the memory consumption seems to increase until all the memory is used up and then it seems to be stuck. Could this be an issue. Can someone give me a hint of what i could look for or change to correct the issues? Command for the run SemiBin2 single_easy_bin --input-fasta Step1_Assembly/bowtie_index/final.contigs.fa --input-bam Step2_binning/inputs_Semibin/contig.mapped.sorted.bam --output Step2_binning/SemiBin_output

Log for semibin run

[2024-01-22 09:22:59,172] INFO: Setting number of CPUs to 40 [2024-01-22 09:22:59,174] INFO: Binning for short_read [2024-01-22 09:22:59,175] INFO: SemiBin will run in self supervised mode [2024-01-22 09:25:47,437] INFO: Binning for short_read [2024-01-22 09:25:47,439] INFO: SemiBin will run in self supervised mode [2024-01-22 09:26:05,682] INFO: Did not detect GPU, using CPU. [2024-01-22 09:26:56,651] INFO: Generating training data... [2024-01-22 09:27:03,431] INFO: Calculating coverage for every sample. [2024-01-22 10:27:06,763] INFO: Processed: Step2_binning/inputs_Semibin/contig.mapped.sorted.bam [2024-01-22 10:36:54,882] INFO: Start training from one sample. [2024-01-22 10:37:13,932] INFO: Training model... [2024-01-22 10:37:18,480] INFO: Generate training data of 0: [2024-01-22 11:33:14,668] INFO: Training finished. [2024-01-22 11:33:15,915] INFO: Start binning.

Output files generated:

contig.mapped.sorted.bam_0_data_cov.csv
data.csv
data_split.csv
model.h5
SemiBinRun.log

Picture showing processes not using CPUs and memory being maxed out. image

HitMonk avatar Jan 24 '24 09:01 HitMonk