When the input fasta file is not wrapped (ie 80 chars per line), the read simulation just hangs but doesn't give an error message..? Assuming this happens for everyone, not just me, would be good to include a quick fasta check and an exit with message if the fasta entry is not wrapped.

Solution: wrap your fasta input file using eg.
seqtk seq -l80 input.fasta

Currently, R packages extracted from their code are reordered so that each package is installed after its direct dependencies.

But this doesn't resolve all problems, currently some packages must be reordered manually.

gen-launchers.sh may fail when used in systems for which /usr/bin/sh is dash and not bash. The function statement (e.g. line 47) seems not to be supported.

Bug

I was trying out tapas and came across the following bug, when running the following I would get the error:

❯ $TAPAS/mapdamagegeomparam --fit-plots fig/fit.png input/mapdamage/misincorporation.txt
Traceback (most recent call last):
  File "/raven/u/jfellowsy/bin/tapas-1.2/scripts/src/mapdamage2geomparam.py", line 436, in <module>
    if __name__=="__main__": main()
  File "/raven/u/jfellowsy/bin/tapas-1.2/scripts/src/mapdamage2geomparam.py", line 68, in main
    A = A.ix[A.Total != 0,:]
  File "/u/jfellowsy/conda-envs/tapas/lib/python3.9/site-packages/pandas/core/generic.py", line 5465, in __getattr__
    return object.__getattribute__(self, name)
AttributeError: 'DataFrame' object has no attribute 'ix'

This is derived from this line:

Where the ix column doesn't exist in the input mapDamage2 misincorporation file.

Workaround

Delete the referred to line in the above script, that refers to ix (assuming your misincorporation file doesn't have any 0s).

I was then successfully able to get the script ot run prpoerly.

"recent" versions of pip do not support the "--system" argument. This makes install_dependencies fail.

I ran through the tutorial and was able to process until the last performance metric step. However, I would like to tune parameters for PE data, not SE data. How could I generate a PE dataset?

Thank you.

In the r_deps list the version of Rcpp is pinned to 0.12.15. This prevents installation with "newer" versions of R (such as 3.5.3 :). I'm not too much into R but if I understood correctly one cannot arbitrarily mix Rcpp and R versions.

The following installation procedure worked for me (providing Rcpp via r-essentials):

git clone https://github.com/mlell/tapas tapas
cd tapas
module load Anaconda/3.7
conda create -n tapas -c bioconda -c r python=3.6 r=3.6 r-essentials
source activate tapas
nano scripts/gen/r_deps.txt  # remove the line containing Rcpp
scripts/gen/install_dependencies
pip install -r scripts/gen/py_requirements.txt --target scripts/lib/python-ext/
bash scripts/gen/gen-launchers.sh