Project_Microbiome

Project repository rs-fMRI of germfree and naïve mice with/without cortical stroke. The aim of this study was to characterize the effect of the gut microbiota in mice with experimental stroke using functional connectivity analysis.

Step-by-Step guide to replicate the analysis

1. Download MRI raw and processed data (raw_data/proc_data) from Edmond – the Open Research Data Repository of the Max Planck Society. The data is grouped, thus GF/SPF, naive/stroke mice are grouped in the correct subfolder already. Further, we provide the AIDAmri_v1.1 and AIDAconnect_v1.0 versions which were used for analysis. This way you save time with pre-set input parameters to reproduce all steps!

2. Install AIDAmri and AIDAconnect. Note! Anaconda Python 3 installation and Matlab_R2018/2019 are required, see the manuals for install instructions.

3. Run AIDAmri pre- and postprocessing steps for T2 and rs-fMRI data, see manual. Note! To save time, you can use batchProc.py. If you do not use the AIDAmri version provided together with the data set, make sure to adjust process_fMRI.py (or your tool to extract the rs-fMRI time series) with the correct repetition time -t 1.42 and the frequency range cutOff_high_sec = 100.0 cutOff_low_sec = 12.5.

4. Extract the graph theoretical measures using AIDAconnect and Matlab

   1. Load the data and generate the graphs mergeDTIdata_input.m
   2. To generate matrices use plotCorrelationMatrices.m (a full list of the atlas labels can be found here)
   3. To generate the histogram plots (frequency of degree/node strength) use plotDistribution.m range 1-50 (strength), 1-100 (degree).
      
   4. To generate the total degree/node strength plots (group difference needs to be calculated separately) use compareTotalStrength.m, compareTotalDegree.m The function was written for multiple time points - you can ignore this as there is only one time point (P0) in this study.
      
   5. To generate the graphs with 20 top connections (ranked according to the Pearson correlation) use analyzeMostConnected.m.
      
   6. To visualize a subnetwork of the whole brain graph use plotSelectedregions.m.
      
   7. To calculate global measures (e.g. density, modularity, etc.) use plotGlobalParameter.m.
      
   8. To calculate the correlation/edge weight between two regions use plotConnectionWeight.m and for all other local measures use plotLocalParameter.m.