add option to make asymmetric nuc substitution matrix

[acvill]

Edited nucleotideSubstitutionMatrix function in pairwiseAlignment.R to include an option to generate an asymmetric nucleotide substitution matrix. This is useful when you don't want to penalize a query sequence that is aligned to a subject / consensus with ambiguous bases.

• For context, see https://github.com/Bioconductor/Biostrings/issues/3#issuecomment-358849578.
• For explanation of implementation, see https://bioinformatics.stackexchange.com/a/19612/3967.

[hpages]

Thanks for the PR. Can you please document the new symmetric argument (in man/substitution_matrices.Rd) and also add some example in the \examples section to illustrate its typical use? Also feel free to add your name in the \author section, with a note about what you contributed to. Something like:

\author{H. Pagès and P. Aboyoun, with contribution from Albert Vill (support
        for asymmetric matrices in \code{nucleotideSubstitutionMatrix()})}

Thanks!

[acvill]

Thanks for the comment, I've implemented the suggested changes to man/substitution_matrices.Rd

[hpages]

Looks good. A few more minor things:

1. Only compute nucleotideSubstitutionMatrix(symmetric = FALSE) once i.e. set the result in a variable (e.g. nsm) and use nsm in your two calls to pairwiseAlignment().
2. Before using nsm, it could be somewhat educational to show some of the values in it e.g.:

   nsm["M", c("A", "C", "G", "T")]
   #  A   C   G   T 
   # 0.5 0.5 0.0 0.0 
   nsm[c("A", "C", "G", "T"), "M"]
   # A C G T 
   # 1 1 0 0 

   and to comment on that.

3. There are too many empty lines between the examples you added and the examples below that. One empty line is enough.

Thanks again, H.

[hpages]

Great, thanks! I'll bump the package version so the new version will propagate to BioC 3.16 (current devel) in the next 24h-48h and become available via BiocManager::install().