Subset subread by single barcode in order to generate inputs for DeepConcensus

[peterdfields]

Operating system

Linux, Ubuntu 18.04.4 LTS

Package name

lima 2.6.0

Using:
  lima      : 2.6.0 (commit v2.6.0)
  pbbam     : 2.1.0 (commit v2.1.0)
  pbcopper  : 2.0.0 (commit v2.0.0-59-g580e770)
  boost     : 1.77
  htslib    : 1.15
  zlib      : 1.2.11

Conda environment

# packages in environment at /home/peter/miniconda3/envs/lima:
#
# Name                    Version                   Build  Channel
lima                      2.6.0                h9ee0642_0    bioconda

My issue here isn't a bug. I may be overlooking what I need to do in the documentation, in which case I apologize for the inconvenience. I'm trying to subset a raw, HiFi barcoded subread bam file to access just subreads related to a single sample. I had initially tried running lima and included only a single barcode. I see the following summary:

ZMWs input                (A) : 4840015
ZMWs above all thresholds (B) : 2834511 (58.56%)
ZMWs below any threshold  (C) : 2005504 (41.44%)

ZMW marginals for (C):
Below min length              : 40 (0.00%)
Below min score               : 69019 (3.44%)
Below min end score           : 69019 (3.44%)
Below min passes              : 0 (0.00%)
Below min score lead          : 69019 (3.44%)
Below min ref span            : 107230 (5.35%)
Without SMRTbell adapter      : 1881585 (93.82%)

ZMWs for (B):
With same pair                : 2834511 (100.00%)
Coefficient of correlation    : 0.00%

ZMWs for (A):
Allow diff pair               : 2353060 (48.62%)
Allow same pair               : 2958430 (61.12%)
Bad adapter impurity          : 93812 (1.94%)

Reads for (B):
Above length                  : 25106741 (99.73%)
Below length                  : 69198 (0.27%)

My interpretation is that the subsetting did not in fact work as expected. I'm trying to create a single subread.bam file so that I can test out the DeepConcensus pipeline on only one of the three samples multiplexed in the Sequel II SmrtCell. Please let me know if any additional information would be helpful.

[armintoepfer]

Perform CCS first and then use all barcodes in the FASTA with lima --split to get one ccs BAM per barcode pair. Before you go into DeepConsensus, actc will only use the subreads from your input ccs reads.

[peterdfields]

Thank you @armintoepfer In the end I'd used lima --split-bam but your solution is probably better.

[amwenger]

We also tested running lima after DeepConsensus, and that produces nearly identical results. So, demux can be run either before or after DeepConsensus, whichever is most convenient.