These python scripts allow metagenomic sequence data to be analyzed with the fast, accurate RNA-Seq abundance estimator kallisto. Both taxa identification and abundance estimation can be performed at the exact-genome level, as demonstrated in our paper "Pseudoalignment for metagenomic read assignment".
##Prerequisites
The full pipeline as shown in our paper requires the installation of genome distance estimator Mash, RNA-Seq abundance estimator kallisto, and python 2.7+.
Python modules needed for running the scripts include NumPy, SciPy, and Biopython.
##Usage
compare_metagenomic_results_to_truth.py compares the output of a range of metagenomic analysis tools such as kallisto to the ground truth of the illumina 100 metagenomic dataset used in our paper. It is run from the command line:
python compare_metagenomic_results_to_truth.py [filename] [program (default kallisto)]
<-g show graphs?> <-s save graphs?> <--taxa level (defaults to all)>
positional arguments:
filename Output file of metagenomic analysis tool
program Source program that created output. Valid options are:
kallisto, kraken, clark, gasic, express. Defaults to
kallisto.
optional arguments:
-h, --help show this help message and exit
-g, --show-graphs Display graphs of calculated errors
-s, --save-graphs Save graphs of calculated errors to file
--taxa TAXA Desired taxa level of analysis. Accepts one of:
strain, species, genus, phylum. Defaults to all
levels.
--dataset DATASET Dataset truth to be compared to. Accepts: i100,
no_truth. Defaults to i100.
--bootstraps BOOTSTRAPS
Directory containing .tsv files for kallisto
bootstraps, to be converted into errors.
plotfunctions.py is a helper module necessary to allow compare_metagenomic_results_to_truth.py to plot the results of analysis.
mash_kallisto_pipeline.py will process the output of Mash (run on a large set of metagenomic reference genomes), and select the top N (user-defined) genomes that match each species that Mash identified in the raw sequenced reads. Python script should be run from the directory containing the raw .fa or .mfa files, and will move matching files to a user-specified directory for later indexing. The script will also process each genome to concatenate contigs/chromosomes, and attempt to look up taxids based on common UIDs present in the file name; a taxid will allow metagenomic programs such as Kraken or CLARK to use these genomes to make a custom database.
mash_kallisto_pipeline.py [-h] [--directory DIRECTORY] [--dry-run] filename top_strains
positional arguments:
filename Mash output file
top_strains How many strains of each species to keep for the
quantification step
optional arguments:
-h, --help show this help message and exit
--directory DIRECTORY
Directory to put files for kallisto index creation.
Default is moving them one directory up.
--dry-run If set, lists files that would be moved, but does not
create or move any files.