DrYak
commented
4 years ago Hello Laura,
However, I tried running this, but I encounter always an error:
/usr/local/bin/b2w: error while loading shared libraries: libhts.so.2: cannot open shared object file: No such file or directory
This error message means that when you try running ShoRAH, it can't manage to find one of its dependencies: the library libHTS (from samtools) against which it was compiled. Did you compile libHTS yourself? Is it also installed in /usr/local/lib? Did you run ldconfig to make your system aware of this new library? If you haven't installed libHTS in a standard location, you might look at either LD_LIBRARY_PATH to specify non-standard locations while running or LD_RUN_PATH/-Wl,-rpath to hard-code non-standard locations when compiling.
Alternatively, I would strongly suggest that you consider installing the bioconda package as conda will automatically handle all dependencies for you and make whole installation experience much simpler. (e.g.: for our V-pipe pipeline, in the tutorials we recommand letting snakemake's built-in bioconda support to handle the package installation).
Additionally, I also was wondering what parameters would be ideal as I don't find a lot of information. Based on a few papers I think that for my purpose following command should work: shorah shotgun -b sample_sort.bam -f sample.fasta -a 0.01 -w 120 -s 3
An example of a current run of ShoRAH in our V-pipe pipeline:
/cluster/work/bewi/pangolin/shorah-test/test/bin/shorah shotgun -t 64 -a 0.1 -w 201 -x 100000 -c 0 -r NC_045512.2:1-29903 -R 42 -b REF_aln.bam -f cohort_consensus.fastaExplanation of options:
-awe're currently using an alpha of 0.1 (which is ShoRAH's default) to control the probability of creating new classes in the dirichlet sampling process. The value of 0.01 that you propose might be helping reduce the false positivies if your sample exhibits low diversity.-wyou must select a width for the window which is divisible by 3 and is shorter that the read length. In our current experiments, we use 250-long reads, so a window of 201 make sure that enough reads will cover a reasonable fraction of the windows (-m, by default 95% fraction of the window). If your experiment uses reads of length 150, using-w 120seems reasonable (and is conveniently divisible by 3).-s 3(we omit it as everything relies on the default 3 overlapping windows in our pipeline).-x 100000we don't really limit the maximum coverage (we can afford the extra processing time).-c 0we currently do not limit the minimum coverage, but beware that benchmarking with simulated synthetic data has show lower quality result in windows with less than 50 coverage. You might consider either filtering at a later point (our current approach) or raising the minimum coverage.-rcan be used to specify a region. More on this later.-Rcan force a random seed if you want more reproducible results.-tthe upcoming 1.99.1 version will allow limiting the concurrent threads. Consider subscribing to our mailing list to be informed about upcoming releases of both our pipeline (V-pipe) and the component we develop such as the present ShoRAH.
however LoFreq doesn't have the ability to construct haplotypes, therefore I turned to this tool.
Please note that ShoRAH 2 has dropped the support for full global haplotype reconstruction (you would need to use ShoRAH 1 for that). It only reconstructs local haplotype within the windows. The .fasta file of this local haplotypes are available in the sub-directory /support. Example from our current research:
# zcat support/w-NC_045512.2-1274-1474.reads-support.fas.gz
>hap_0|posterior=1 ave_reads=404.707
GTTAAAGCCACTTGCGAATTTTGTGGCACTGAGAATTTGACTAAAGAAGGTGCCACTACTTGTGGTTACTTACCCCAAAATGCTGTTGTTAAAATTTATTGTCCAGC {...}
>hap_1|posterior=0.0254284 ave_reads=0.0254284
GTTAAAGCCACTTGCGAATTTTGTGGCACTGAGAATTTGACTACAGAAGGTGCCACTCCTTGTGGTTACTTACCCCAAAATGCTGTTGTTAAAATTTATTGTCCAGC {...}For each reconstructed local haplotype you get:
- an arbitrary
hap_identifier (the numbers are arbitrary and in rough order of posterior. They do not match across windows) - the
posteriorprobability of this haplotype after sampling. I would suggest filtering and keeping only high probability posteriors (currently, when calling SNVs, ShoRAH only keeps haplotypes with posteriors > .9) ave_readsthe number of reads which contributed to this haplotype during the considered iterations. (you could also consider filtering that)
In the above example, I would definitely trust hap_0 (has a high probability, and is based on a large amount of reads), but not hap_1 (low posterior probability, few reads contributing to it).
You could also have a look at the discussion in issue #66
Last important note:
As it turns out, until now we have mostly worked with viruses which have a single continuous genome: HIV, HCV and SARS-CoV-2.
I haven't tested how the code reacts in the case of a virus with a segmented genome as the orthomyxoviridae such as Influenza.
I would be very interested if you could report any error, so I can try to fix segmented viruses support in b2w (or elsewhere inside ShoRAH 2).
In case of errors, a possible circumvention would be to force the processing of specific regions using the -r parameter.
Please keep me informed of your successes.
LauraVP1994
(It works best by selecting one of the discrete colours from the palette drop down)
Dear,
I am very interest in using your tool for my project. I'm working on influenza reads that were sequenced with Illumina MiSeq. I was able to extract quasispecies with LoFreq, however LoFreq doesn't have the ability to construct haplotypes, therefore I turned to this tool. However, I tried running this, but I encounter always an error:
/usr/local/bin/b2w: error while loading shared libraries: libhts.so.2: cannot open shared object file: No such file or directory
I also tried this with the amplicon example that was given, but same error was given. I don't know if you have an idea what could be the cause of this?
Additionally, I also was wondering what parameters would be ideal as I don't find a lot of information. Based on a few papers I think that for my purpose following command should work: shorah shotgun -b sample_sort.bam -f sample.fasta -a 0.01 -w 120 -s 3
However, I don't know if the other parameters should be set or if the default is sufficient?
Kind regards and thanks in advance Laura