deepconsensus-0.2.0 uses a lot of system RAM, so maybe best to split jobs into pieces, example provided about deepconsensus HOT 4 CLOSED

Parallelizing DeepConsensus is actually easier than that. We plan to write a full guide, but here is a short summary:
There is a --chunk option in ccs as mentioned in the quick start, which will produce sharded outputs -- we usually do 500 shards for a full SMRT-cell. There is no need to samtools sort any of the bam files since they are already in sorted order by ZMW. This order MUST be the same between the subreads_to_ccs.bam and ccs.fasta files, which they are coming out of the ccs and actc steps, so it's best not to run anything like samtools sort that might change that order. Beyond that, you can follow the rest of the quick start separately for each shard.

from deepconsensus.

Ok, I will try the --chunk option without samtools sort now.

from deepconsensus.

Update: I just released a major change to the DeepConsensus quick start with detailed guidance for parallelization across multiple machines: https://github.com/google/deepconsensus/blob/r0.2/docs/quick_start.md

from deepconsensus.

maybe just tp add here: i tried to follow the guide and created a snakemake-based workflow for this:https://github.com/WestGermanGenomeCenter/deep_snake

from deepconsensus.