parse_kraken2_report

Taxonomically informed parsing of kraken2 report output.

Motivation

I find simple parsing of kraken2 output report can be a bit annoying. If you parse only using the percentage/number of reads assigned, then you end up not only with Salmonella enterica but also everything between Salmonella enterica and root. If you filter to only include species or genus, then you could end up missing something significant at a higher level.

This script will read in a kraken2 report, establish the taxonomic relationships between the results, and then print out as many or as few taxonomic levels as you like, starting from the "tips".

Note

2020-10-01 This software is fresh out of the box. Please sanity check results and report any bugs.

Arguments

required named arguments:
  -i INHANDLE           Path to kraken report file (default: None)

optional arguments:
  -h, --help            show this help message and exit
  -n SAMPLE_NAME        Sample name to be included in output. If not give,
                        will take everything before the first period of the
                        file name. (default: None)
  -l NUMBER_OF_LEVELS   How far up from each tip do you want to check? If not
                        working as expected, you may ned to alter -r as well.
                        (default: 2)
  -p PERCENT_READS_ASSIGNED_THRESHOLD
                        Minimum threshold of percent_reads_assigned for
                        reporting (default: 0.05)
  -r TAXONOMIC_RANKS    Taxonomic ranks which you want to report given in
                        comma-separated, upper-case format, no spaces. Rank
                        codes should reflect Kraken2 output documented here ht
                        tps://github.com/DerrickWood/kraken2/wiki/Manual#sampl
                        e-report-output-format. Don't include number
                        indicating sub-ranks. If not working as expected, you
                        may need to alter -l as well. (default: S,G)
  -t                    Include this option if you want to print the tree
                        (default: False)

Usage examples

Defaults:

python kraken2_scripts/parse_kraken2_report.py -i example.kraken_report.txt -n sample1

Report species, genus and family (note the -l is set to 3 as well).

python kraken2_scripts/parse_kraken2_report.py -i example.kraken_report.txt -l 3 -r S,G,F

This is quite a good option (considering switching the default), print all the "tips" which meet the percentage of reads mapped criteria.

parse_kraken2_report.py -l 1 -r S,G,F,O,C,P,K,D -i example.kraken_report.txt

Dependencies

Uses the python package anytree which can be installed from pip. Developed using anytree v2.8.0.