SamestrGal: Command-Line Reproduction and Benchmarking

This repository contains the steps and commands required to run SamestrGal's tools from the command line, and to reproduce the benchmarking of SamestrGal against the SameStr Nextflow workflow. Additional files are provided to reproduce the taxonomic graphic in the thesis. Links to the datasets used in both workflows, as well as the results obtained, are also given.

Repository Structure

.
├── scripts/
│   ├── nextflow_files/        # Corrected samestr.nf and params.yml for the Nextflow workflow
│   └── taxonomic_graphic/     # Script to reproduce the taxonomic graphic
└── terminal_commands/
    ├── testing_samestr_tools.sh   # Commands to run SameStr tools from the command line
    └── nextflow_workflow.sh       # Commands to run the SameStr Nextflow workflow

Replication of benchmarking

The links to the input datasets used for benchmarking can be found in nextflow_workflow.sh. These datasets can also be used for testing the individual tools in the command line. For this, the database downloads and command instructions can be found in testing_samestr_tools.sh.

For the benchmarking, run all datasets in Nextflow following the steps in nextflow_workflow.sh. It should take approx. 3-4h for each case. The most relevant files for the benchmarking will be in results/sstr_summarize/.

For SamestrGal, create 5 new histories, one for each case. The Galaxy histories from the thesis can be found here:

Dataset	History
FRICKE Case 16	Galaxy History
FRICKE Case 19	Galaxy History
FRICKE Case 20	Galaxy History
FRICKE Case 28	Galaxy History
FRICKE Case 54	Galaxy History

In order to compare the results generated by both workflows, the first file to check is results/sstr_summarize/taxon_counts.tsv and Taxon counts in SamestrGal. From here, the amount of taxa detected by the workflow at each taxonomic rank in each sample can be compared. Then the actual shared taxa between samples can be analysed in the results/sstr_summarize/sstr_cooccurrences.tsv table in Nextflow and Strain co-occurrences dataset in SamestrGal. This table shows the shared taxa between all samples in the case. The relevant ones are the comparison between the Pre-FMT and Post-FMT sample, which shows the taxa the patient has kept from before the FMT treatment to after, and the comparison between Donor and Post-FMT sample, which shows the taxa that were engrafted from the healthy donor microbiota into the patient after FMT treatment was completed. This is a direct way of seeing how similarly the shared taxa were detected in both workflows. In order to see the exact clades that were detected as shared strains in the co-occurrences dataset, the file results/sstr_summarize/sstr_strain_events.tsv in Nextflow and Strain events in SamestrGal can be analysed. There the clades can be analysed one by one. The overlap and MVS score are also included, which can then help understand why a clade was detected as a shared strain in one workflow and not in the other.

The Excel files containing the comparison of both workflows' results can be found in the following link:

https://doi.org/10.5281/zenodo.21630816

Reproducing the Taxonomic Graphic from the SameStr publication

This graphic reuses the plotting notebook from the original SameStr publication's repository:

https://github.com/danielpodlesny/fmt_rcdi

Steps

1. Clone the original repository:

   git clone https://github.com/danielpodlesny/fmt_rcdi.git

2. Download data.zip from the repository and take the files microbe.taxonomy and samples.case_summary: https://github.com/danielpodlesny/fmt_rcdi/blob/master/data.zip

3. Create new versions of these two files using the data obtained from this thesis's SamestrGal results. For the samples.case_summary the first column indicates whether the data is a Control sample or a rCDI sample. Copy the Control sample rows unchanged from the original file into the new samples.case_summary file. For the rest:

   • New microbe.taxonomy: created by joining all species with detected abundance for all 5 rCDI cases, which can be obtained from each sample's MetaPhlAn taxonomic profiles. In SamestrGal the dataset is called: Predicted taxon relative abundances.

   • New samples.case_summary: the rCDI case rows are generated from several SamestrGal outputs. For each species-specific SNV profile detected in SameStr Convert for all samples of the five cases, a row has to be added along with metadata such as the sample name and the case it belongs to. The nucleotide coverage is also needed for all three FMT samples (Pre-FMT, Donor, and Post-FMT) and can be taken from the per-sample alignment statistics generated by SameStr Convert.The relative abundances also need to be added for each clade in each sample, and can be obtained from the MetaPhlAn taxonomic profile. For each row, the overlap and MVS values generated by each pairwise comparison that SameStr Compare generates also need to be added in different columns. The column names for it are:

     • Pre-FMT/Donor: for overlap and MVS score values between the patient before FMT treatment (Pre-FMT) and the Donor.
     • Pre-FMT/Post-FMT: for overlap and MVS score values between the patient before FMT treatment (Pre-FMT) and the same patient after treatment (Post-FMT).
     • Donor/Post-FMT: for overlap and MVS score values between the Donor and the patient after treatment (Post-FMT).

Study_Type	Study	Case_Name	Sample.recipient	Sample.post	Sample.donor	Days_Since_FMT.post	fmt_success	fmt_success.label	Pre-FMT/Donor.overlap	Pre-FMT/Post-FMT.overlap	Donor/Post-FMT.overlap	Pre-FMT/Donor.mvs	Pre-FMT/Post-FMT.mvs	Donor/Post-FMT.mvs	rel_abund.recipient	rel_abund.post	rel_abund.donor	event	source	analysis_level	species	n_covered.recipient	n_covered.post	n_covered.donor
rCDI	FRICKE	FRICKE_Case_20	FRICKE_20A.pair	FRICKE_20E.pair	FRICKE_20B.pair	84	TRUE	Resolved	8086	8079	163998	0.99381647	0.98799356	0.99986585	0.023261	2.383998	3.04135	replaced	donor	strain	Agathobacter_rectalis	8180	171969	172129

Besides the data obtained from the results, an analysis level, a source and an event need to be added. The exact conditions that were used in the taxonomic graphic in SameStr were not stated. Based on all samples that were analysed, this is the logic that was abstracted from it:

The nodes represent all possible events that can be added to a row based on certain conditions. For those events in green, the source is donor. For those in blue, the source is self. For those in yellow, the source is unique. For those in purple the source is both. There are also two events without a colour, those are source same_species, but the script doesn't take them into account when creating the figure. The analysis level can also be determined by the colours in the image. For each SNV profile row in samples.case_summary, first check if there is abundance in all 3 samples (rel_abund.recipient, rel_abund.post, rel_abund.donor), here recipient refers to the Pre-FMT sample. To check the overlap of that SNV profile, check columns (Pre-FMT/Donor.overlap, Pre-FMT/Post-FMT.overlap, Donor/Post-FMT.overlap). For the decision node called All overlap > 5000 here is meant that all 3 overlap columns have overlap values that are higher than 5000. Similarly, in How many MVS > 0.999 refers to how many of the MVS columns (Pre-FMT/Donor.mvs, Pre-FMT/Post-FMT.mvs, Donor/Post-FMT.mvs) are over the 0.999 threshold.

4. Place the new microbe.taxonomy and samples.case_summary files in fmt_rcdi/data/ to replace the originals.

5. Run the notebook that generates the figure:

   Rscript -e 'rmarkdown::render("notebooks/pub.Figure_5.Rmd")'

   The script used is the original pub.Figure_5.Rmd with two changed lines: lines 468 and 478 use filter(rCDI >= 5, Control >= 3) %>% instead of filter(rCDI >= 10, Control >= 5) to represent that less data was used in this analysis than to generate the original plot. The exact script used for SamestrGal can also be found in taxonomic_graphic/.

Data Table Entries

The exact rows added for the rCDI cases with the data from SamestrGal used in the thesis to generate the plot can be found in the following link:

https://doi.org/10.5281/zenodo.21642582