Morning. Finally I needed to include the `--no-pre...
# phyloplace
a
Morning. Finally I needed to include the
--no-pre-mask
flag since it happened for more than 10% of the samples was too much. The options are 1) to include the HMM cleaning as you said and 2) to add the flag, which works because I tested it with one sample. Given that these are a minority of reads I think its a good workaround, and sometimes the users won't have an specific HMM to use I think.This opened though a question when I tried to run the workflow adding by myself the tag. I thought that I should add it in the
nextflow.config
following this schema:
Copy code
process {
    withName: 'EPANG_PLACE' {
        ext.args = ['--no-pre-mask']
    }
}
But if I add this, the software is unable to find the files, throwing this error.
ERROR ~ Argument of file function cannot be null
I am guessing that the call
ext.args
does not include the value upfront but substitutes partially the input that you give manually? This is my call:
Copy code
nextflow run nf-core/phyloplace     --input data/placement_phyloplace/SIT/SIT_undone_samsheet.csv     -profile singularity --outdir data/placement_phyloplace/SIT
Thanks! PD: it is possible that the error comes from me using a previous version of nextflow (v23) or from the pipeline, I will be testing this, but if your intuition tells you that this is the problem just say that and sorry beforehand!)
d
I think you need to do this:
Copy code
withName: 'EPANG_PLACE' {
        ext.args   = { "--model ${meta.model} --no-pre-mask" }
    }
✅ 1
a
It seems that it also happens with the new version of the pipeline:
Copy code
>000001|MATOU-v1.5.731
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
-----------------M------------------------------------------
------------------------------------------------------------
---------------------------------------------------
>000002|MATOU-v1.5.731
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
--------------------------------RFGQGNH---------------------
---------------------------------------------------
>000003|MATOU-v1.5.731
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
-------------------------GKRTRVFFPSSE-----------------------
------------------------------------------------------------
---------------------------------------------------
>000004|MATOU-v1.5.731
------------------------------------------------------------
------------------------------------------------------------
----------------E-------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
---------------------------------------------------
>000005|MATOU-v1.5.3120
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
-------------S----------------------------------------------
------------------------------------------------------------
---------------------------------------------------
>000006|MATOU-v1.5.3120
------------------------------------------------------------
-----------------------------V------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
---------------------------------------------------
Perhaps there is an option for the hmm in which we can filter out the aligned queries that look like this? Since these are a lot it would speed up the whole process. I still do not understand why the premasking does not work !
d
Easiest is probably to pass a param for minimum score or maximum e-value to
hmmsearch
via
ext.args
in a config