0

Something is wrong with your input data.

Here is an error log for the same: Error Log

Microbial association networks were generated using the data available from the LogMPie project which aims to create a pan-India profile of the human gut microbiome using 1004 subjects. For this case study, we focused our analysis on the four broad Indian regions namely North, South, West and East having approximately equal distribution of the samples.

Time-series human gene-expression data post-infection by two strains of Mycobacterium tuberculosis (H37Ra and H37Rv), overlaid on the human protein-protein interaction network was used to generate multiple gene perturbation networks. For this case study we considered the networks generated using three post infection time points (16h, 48h and 90h) corresponding to the two strains resulting in a total of six networks.

iVikodak was utilized for inferring as well as analyzing 91 publicly available 16S metagenomic datasets corresponding to sub-gingival microbiome samples from 30 Periodontally healthy control subjects (HC), and sub-gingival samples from 30 shallow (PS) and 29 deep (PD) periodontal pockets of subjects with Periodontitis. Edge lists pertaining to functional driven co-contribution networks, as available in iVikodak Global Mapper demos, were utilized for this example.

In order to demonstrate the utility of NetConfer tool, we chose two publicly available available research reports and their corresponding data.

Case Study 1: First case pertains to gut microbiomes from 60 multiple sclerosis patients spanning three sets (28 without any treatment, 18 treated with Interferon & 14 treated with Copaxone for at least six months) and 43 control subjects. More details of the case study, along with links to downloadable edge-lists may be accessed below.

Case Study 2: Time-series human gene-expression data post-infection by two strains of Mycobacterium tuberculosis (H37Ra and H37Rv), overlaid on the human protein-protein interaction network was used to generate multiple gene perturbation networks. For this case study we considered the networks generated using three post infection time points (16h, 48h and 90h) corresponding to the two strains resulting in a total of six networks.

Microbial association networks (at genera level) for the microbiome datasets corresponding to all the four groups (‘Control’, ‘Untreated’, ‘Interferon’ and ‘Copaxone’) were first generated using the CCREPE tool. CCREPE calculates the statistical significance of a correlation (Spearman in this case) between a given pair of genera using a re-sampling and null distributions of correlation values (obtained by 1,000 iterations) and assigns and edge between a pair of genera for every predicted significant positive correlation (p value less than 0.005).

The observations from the microbiome networks of the case study presented in the manuscript clearly indicated that the microbiome is more affected during the treatment as compared to the disease itself. Given the availability of the host gene perturbation network, it was imperative to compare the effect of the treatment (copaxone and interferon) against the untreated group from a network perspective.

In order to understand host response in disease state as well as post treatment, the most perturbed genes in the disease and treatment groups (having a median expression fold change >= 1.5) with respect to the control group were identified from the gene expression data. The resultant list of most perturbed genes was then overlaid on the STRINGS Human PPI (protein-protein interaction) network using the NetworkAnalyst tool to extract the network of connected perturbation. A total of six gene perturbation networks, two each (for T-cells and monocytes) in the untreated, interferon (treated) and copaxone (treated) patients were generated.

The expression data corresponds to a selected 568 immune related host genes (derived from CD4+ T-cells and CD14+ monocytes) from a subset (18 multiple sclerosis and 18 healthy subjects) of the above mentioned groups of samples, and generated PPI networks using highly perturbed genes by overlaying them on STRINGS Human PPI global network.

Five simplified hypothetical networks are created to demonstrate the utilities of NetConfer. These are tab separated edge lists representing five different weighted networks. Download the edle-lists using the buttons below: