conda environment and test difference

[dkoppstein]

Hi, I've been able to successfully use the docker image to run the BRACER test on my linux box and get the expected results. Unfortunately, our HPC does not support Docker. Thus, I've been trying to recreate the dependencies using bioconda. I've tried to match the versions as closely as possible -- see the conda environment file below. I've also recreated the gencode v27 index and ran the python script to reformat it. However, strangely, I consistently get only 6 reconstructed BCRs when using this environment, not 7. I've attached the STDOUT and STDERR logfiles and the conda environment file I'm using, as well as the contents of the results directory. It seems like Trinity might be failing for one of the BCRs, but I can't tell any more from the log files. Is there any chance you could support using conda for those of us who are not fortunate to have Docker availability on our HPCs?

Best, David

---

environment-explicit.txt STDIN.e1623959.txt STDIN.o1623959.txt results.tar.gz

[scharch]

I'm having the same issue: the heavy chain for cell1 is not detected. It looks like they are being discarded in the second round of alignment, note the second half of the output below:

Short read length detected. BraCeR will run two rounds of alignment for heavy chain

##BCR_H##
56438 reads; of these:
  56438 (100.00%) were paired; of these:
    35350 (62.64%) aligned concordantly 0 times
    21088 (37.36%) aligned concordantly exactly 1 time
    0 (0.00%) aligned concordantly >1 times
    ----
    35350 pairs aligned concordantly 0 times; of these:
      73 (0.21%) aligned discordantly 1 time
    ----
    35277 pairs aligned 0 times concordantly or discordantly; of these:
      70554 mates make up the pairs; of these:
        69268 (98.18%) aligned 0 times
        232 (0.33%) aligned exactly 1 time
        1054 (1.49%) aligned >1 times
38.63% overall alignment rate
56438 reads; of these:
  56438 (100.00%) were paired; of these:
    56438 (100.00%) aligned concordantly 0 times
    0 (0.00%) aligned concordantly exactly 1 time
    0 (0.00%) aligned concordantly >1 times
    ----
    56438 pairs aligned concordantly 0 times; of these:
      0 (0.00%) aligned discordantly 1 time
    ----
    56438 pairs aligned 0 times concordantly or discordantly; of these:
      112876 mates make up the pairs; of these:
        112876 (100.00%) aligned 0 times
        0 (0.00%) aligned exactly 1 time
        0 (0.00%) aligned >1 times
0.00% overall alignment rate

Unfortunately, I don't see a way to override this, so I can't test.

[idalind]

Hi both, thanks a lot for your feedback and apologies for the late reply. I will look into this issue over Christmas and see if I can sort it out. Regarding conda we are not currently working on this, but I agree it would be very useful and we will look into it!

[scharch]

Interestingly, this seems to work with bowtie 2.2.5 (the output I pasted above was with bowtie 2.3.3.1). However, Trinity fails with the older version of bowtie2-build because it apparently doesn't accept the --threads option as Trinity expects.

[idalind]

Hi both. Again, apologies for the late reply. Could you please try with bowtie 2.2.8? This is the version I personally use, without the Docker image. If this solves the issue, I will figure out how to make BraCeR compatible with newer bowtie 2 versions.

[scharch]

Yes, it works properly with bowtie 2.2.8

[idalind]

Hi @dkoppstein, did you get BraCeR working using bowtie 2.2.8?

[dkoppstein]

Hi @idalind, I confirm that Bowtie 2.2.8 works for me. Here is my conda environment.

[idalind]

Thanks a lot for the feedback @dkoppstein ! I'm closing this issue now.

[Zhemm]

I’m looking for a large database for heavy chain and light chain of antibodies. Does anybody has any suggestions?