Script description

Package requirement

PySEAT may have conflict with numpy version. We recommand: numpy = 1.22.4 and pyseat = 0.0.1.3

To reproduce the result, first please download and unzip data.zip to /data directory.

If you want to start the analysis from GCN, please run /script_procedure/step1_compute_distance first to compute distance matrix of GCN. This may take some time. To save time, you can directly use sp_d.tsv in /data directory which is preproduced by step1_compute_distance.

Prior structure is generated by scripts in /script_GCN_d3/GCN_tree, please run this script before FMT, NSCLC, Anti_analysis which depend on the prior sturcture.

input: following files (can be found at data/)

 metadata.tsv(metadata from GutMeta, disease should be in the header for phenotype comparison)
 abd.tsv (header: species, index: sample name)
read by abd_profile.input_profile()
 GCN.tsv (header: KO, index: species)
read by GCN.input_GCN()

script of taxonomy abundance difference (script_abundance_check)

Check the abundance difference for each taxon (including NAFLD 16s OTU).

script of completeness (script_completeness)

1. completeness
   Compute the module completeness of each taxon (including NAFLD 16s OTU).

2. test_diff
   Test completeness enrichment based one the GCN prior GCN structure and should be used after GCN_tree script in script_GCN_d3

script of prior GCN structure (script_GCN_d3)

1. GCN_tree
   Make prior GCN tree structure.

2. SE_diff / NFR_diff
   Check SE/nFR difference of disease and health group.

3. distribution_se
   Plot SE distribution for disease and health group.

script of FMT (FMT)

GCN_tree result is required

Scripts related to two FMT dataset analysis.

1. analysis_se
   Mutiple regression on SE value, days after FMT and fraction at each cluster/super-cluster.

2. analysis_nfr
   Mutiple regression on nFR, days after FMT and fraction at each cluster/super-cluster.

script of Antibiotic treatment (script_Anti_analysis)

GCN_tree result is required

1. analysis_se/analysis_nfr
   Check SE/nFR difference of control and exposed group at each clsuter/super-cluter.

2. analysis_se_exposed/analysis_nfr_exposed
   Check SE/nFR difference of six participants exhibited a bloom of the opportunistic pathogen Enterobacter cloacae complex at the E7 timepoint in exposed group and control group at each clsuter/super-cluter.

3. merge
   Merge and plot the difference test result of nFR and SE in control and exposed group.

4. merge_exposed
   Merge and plot the difference test result of nFR and SE of the six samples and control group.

5. boxplot Draw boxplot for SE at each cluster/super-cluster.

script of lCFR procedure （script_procedure）

1. step0_NAFLD
   An example of comparing keyston clusters of taxa on NAFLD dataset.

2. step1_compute_distance
   An example of computing taxa distance and KO distance from GCN.

3. step2_cluster_analysis
   An example of analyzing keystone cluster and keystone taxon for metagenomics abundance profiles in cMD by constructing posterior structure.

4. step3_count_support
   An example of checking valid keystone-taxon enterotype with more than one network of size larger than 10 supporting.

5. utils

   a. log_effect
   An example of computing lCFR and showing distribution of lCFR values and CFR values without log and normalization.

   b. nestedness_experiment
   An example to test the nestedness compared with NULL experiments of lCFR.

   c. evaluation
   An example to evaluate the feature of GCN.

6. draw_*
   Scripts used to plot the result of previous step.

script of NSCLC (script_NSCLC)

1. SE
   Test difference of SE between response group and non-response group at each cluster/super-cluster and compute FR S score for each sample.

2. sig_SE
   Test difference of SE between response group and non-response group at SIG1/SIG2 clsuter raised in original study and compute S score for each sample.

3. distribution Plot SE distribution for response group and non-response group.

4. combination
   Compute combined S score for each sample.

5. The r script Used to produce the analysis in original study and is provided by https://github.com/valerioiebba/TOPOSCORE/tree/main.

plot keystone graph (script_keystone_graph)

GCN_tree result is required
abundance difference result is required

1. run.ipynb
   Plot keystone result.

script of eigenspecies analysis (script_eigen_graph_preservation)

GCN_tree result is required

1. run.ipynb
   Find eigen species and plot the result.

script to predict disease (script_predict)

GCN_tree result and SE values are required

1. CRC_recurrent_ROC.ipynb
   Predict CRC.

2. IBD_ROC.ipynb
   Predict IBD.