Database processing update

[farchaab]

PR for version 0.9.20

Summary

• Updated pre-processing for Silva 138.1 (with species)
• Updated pre-processing for Greengenes2
• Updated rdp_classifier version from 2.5 to 2.14
• Optimized lineage2taxTrain.py
• Added taxa formatting when no pre-processing for Silva 138.1, Greengenes2, ezBiocloud and UNITE
  • Solved convergent taxa in Silva 138.1 and Greengenes2

Details

Additions

• 5627a03 Added conda env and container for pandas
• d1d3860 Added python script for derep_and_merge_taxonomy:
  • Fills empty taxa with placeholder names (ex: Escherichia-shigella_s)
  • Replaces Incertae, unkown and endosymbionts with NaN (caused convergent taxons for rdp)
  • For unclassified taxa, each taxa corresponding to a different cluster will have a unique index
• cc8540a Added db_version parameter
• 6f30c51 Added script to clean taxonomy files when no processing
• d574275 Added list of classifiers for preparing databases (RDP, QIIME, and DADA2 by default)
• 1619919 Added RDP memory as parameter

Changes

• 4a94081 R scripts linting
• bdb8a0e Changed R to python script for derep and merge taxonomy
• c79db5a a467bc7 Updated python env and container to 3.12.3
• 991605e Replaced numbers_species and numbers_genus by a dict {'Species': 2, 'Genus': 4}
• 9920ea4 Use rdp_classifier instead of rdptools
• 7c32765 Updated tax_formatting.py to work with Greengenes2

Fixes

• d1d3860 for Database processing with latest SILVA and greengenes versions #42
• 145cfd0 Fixed lineage2taxTrain.py slowness by replacing list to set (thanks @tpillone !)
• 35b648c Fixed some convergent taxa by propagating NaN to downstream ranks