A MATLAB package to quantify the functional redundancy of human microbial community

Liang Tian^1,2, Xu-Wen Wang¹, Ang-Kun Wu^1,3, Yuhang Fan^1,4, Jonathan Friedman⁵, Amber Dahlin¹, Matthew K. Waldor^6,7, George M. Weinstock⁸, Scott T. Weiss¹ & Yang-Yu Liu¹

¹Channing Division of Network Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts 02115, USA
²Department of Physics, Hong Kong Baptist University, Hong Kong, China
³Department of Physics and Astronomy, Rutgers University, Piscataway, New Jersey 08854, USA
⁴Department of Bioengineering, Stanford University, Stanford, California 94305, USA
⁵Department of Plant Pathology and Microbiology, Faculty of Agriculture, Food and Environment, The Hebrew University of Jerusalem, Jerusalem, Israel
⁶Division of Infectious Diseases, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts 02115, USA
⁷Howard Hughes Medical Institute, Boston, Massachusetts 02115, USA
⁸The Jackson Laboratory for Genomic Medicine, Farmington, Connecticut 06117, USA

Last update: Sep 06, 2020

The files contain:

1. Abundance table and genomic content network for Stool sample of HMP (HMP_Stool.mat): Table of abundance at stain level of 553 stool samples from the HMP (http://www.hmpdacc.org/ ).
2. Matlab scripts (script_FR_GCN_randomization.m & script_FR_otu_randomization): The matlab code imports the abundance table and genomic content network, calculates the functional redundancy for the real abundance and genomic content network and also for the randomized counterpart, and then plot the figures.
3. Matlab script (Genomo_evolution_model.m) of a genome evolution model. In this script, we use the Nestedness Temperature Calculator to quantify and visualize the nested structure of genomic content network, which is based on the BINMATNESS algorithm (https://bimat.github.io/alg/nestedness.html).
4. Matlab functions used by the scripts.

Running the tutorial:

1. Download all the files in this repository to a local directory.
2. Run the Matlab file. Running times for both of the scripts are less than 1 min. (The code was written on MATLAB R2016b)

Output:

1. Matlab figure showing normalized functional redundancy of real abundance and genomic content network compared with that for randomized genomic content network (left) and for randomized abundance table (right).
2. Mann-Whitney U test p values and adjusted q values (Benjamini-Hochberg).