bigSMALL v1.3
Released: 12/1/2016
Updated: 4/7/2017
by: Matthew L. Jenior
Department of Microbiology & Immunology University of Michigan mljenior@umich.edu
When using, please cite: Jenior, M.L., Leslie, J.L., Young, V.B., Schloss, P.D. (2017). Clostridium difficile colonizes alternative nutrient niches during infection across distinct murine gut communities. mSystems. 2 (4); e00063-17.
Distributed under the GNU General Public License
#---------------------------------------------------------------------------#
Python scripts and workflow by Matthew Jenior, University of Michigan, 2014 - 2016
#---------------------------------------------------------------------------#
The function of this package is to infer putative metabolites most likely acquired for the environment based on transcriptomic data mapped to KEGG orthologs. Bigsmall generates a bipartite metabolic based on the reaction data associated with each KEGG ortholog and then integrates transcript abundances to predict demand for metabolites based on the transciption of adjacent enzyme nodes. Monte Carlo simulation is also applied to create a standard of comparison that reflects random noise.
KEGG reference files and reference creation script can be found in the support directory
Examples of input files for each program can be found in the examples directory
#---------------------------------------------------------------------------#
Calculates relative importance of a given metabolite based on the expression of surrounding enzymes in a metabolic network
python bipartite_graph.py ko_expression.list
Positional, required argument:
expression_file - two column file of KEGG ID and transcript abundance
Optional arguments:
-h, --help - show this help message and exit
--name - Organism or other name for KO+expression file (default is organism)
--iters - iterations for random distribution subsampling (default is 1000)
#---------------------------------------------------------------------------#
Multi-level inference of substrate competition and cooperation between transcriptome-informed genome-scale models
Calculates putative community-level and pair-wise metabolic interactions between species from aggregated bigSMALL analysis
python crosstalk.py [-h] bigSMALL_interactors [--p_value P_VALUE]
Positional, required argument:
bigSMALL_interactors 1 column file of species with bigSMALL output for each (directories)
Optional arguments:
-h, --help show this help message and exit
--p Minimum p-value for metabolites to be considered in calculations
--norm Option to normalize scores from each metabolic network based on their individual sequencing coverage
#---------------------------------------------------------------------------#
Sample files to be used as examples with each of the respective scripts
ko_expression.tsv - list of KO codes and corresponding expression values for Clostridium difficile strain 630
ko_expression.bipartite.files - Output of bipartite_graph.py for Clostridium difficile 630 during a mouse infection
interactions.files - Text file that lists the memebrs of the community you wish to model metabolic interaction and have bigSMALL output directories
community.files - Output directory of all pair-wise comparisons of species in the community analyzed as well as overall metrics of community demand for metabolites
#---------------------------------------------------------------------------#
McGill, R., Tukey, J. W. & Larsen, W. A. (1978). Variations of Box Plots. The American Statistician 32, 12–16.
Ogata, H. et al. (1999). KEGG: Kyoto encyclopedia of genes and genomes. 27, 29–34.
Patil, K. R. & Nielsen, J. (2004). Uncovering transcriptional regulation of metabolism by using metabolic network topology. PNAS 102 (8), 2685–2689.