Urgently needed modifications

[ImagoXV]

List of needed features and modifications

• [x] Personal database implementation https://github.com/ImagoXV/NanoASV/commit/a48298e33ebc8c7b62db2f117e6a6d5e207543f7 - Personal database OK. Indexing step at first NanoASV run
• [x] Change the moment when the database is indexed. Need to write on computer outside of container https://github.com/ImagoXV/NanoASV/commit/9a1bdec48ac0f3c1ef484af6e023d6d248a9a9fd - SILVA is no longer indexed at building step
• [x] Rewrite samtools filtering process. Remove bash language to let samtools do the job. https://github.com/ImagoXV/NanoASV/commit/427cfc05db14844528eeb4921a5549b496196d38 - Samtools does a proper filtering job now without parsing sam files with grep. I wonder if indexing step is usefull or not
• [x] Make installation process more user friendly Should be handled quite well with uploading a singularity image file (SIF) executable as release. No more building from source with docker !
• [ ] Make more comparison with Nyygaard pipeline
• [ ] Run published data with NanoASV and compare composition dataset for example

Additional fixes or changes:

• [x] update to Silva 138.2,
• [x] fix indentation throughout script.sh,
• [x] eliminate intermediate or temporary files as soon as possible,
• [x] remove all commented code,
• [x] replace paths such as /opt/ with a variable (easier to modify),
• [ ] put DEBUG code behind a test: [[ ${DEBUG} == "TRUE" ]] && { code ;}
• [ ] in the bash script, is there an alternative to declaring env variables for parallel?
• [x] in R, add package names to function that are not R-base (e.g.; dplyr::mget()) to make it easier to understand the origin of each function and the role of each package

[frederic-mahe]

Personal database: you should document somewhere the exact format required by your pipeline. You should also set up basic tests to check if the database provided by the user is ok before running with it.

[ImagoXV]

Yes that's a excellent point. It will need some work

Notably, Taxonomy is extracted from the fasta file. It is relevant for Silva or GreenGenes, but needs a lot of conditions if a regular fasta file is loaded At this point the only real test and homogeneization is the automatic transformation from a multiple line to a single line fasta file

[frederic-mahe]

I've made a list of suggestions after reading your recent commits.

[ImagoXV]

Thank you very much Fred, I'll dive into it