A Proposed Workflow to Robustly Analyze Bacterial Transcripts in RNAseq Data from Extracellular Vesicles

This project was created to analyze bacterial (or non-human in general) transcripts from RNA-seq data. The project is organized as follows:

data/                                             | .fastq files (raw or processed and relevant) - The FASTQs used for the project are available.
database/                                         | Files necessary for database creation.
results/                                          | Directory where (1) profiling outputs and (2) results from the analysis are stored.
|
-- counts/                                        | It contains the number of reads mapped/unmapped in each iteration of the host-mapping and profiling process.
                                                    These files are generated by notebook `0_get_fastq_read_counts.ipynb` but require heavy profiling output files that cannot be stored. 
-- differencial_abundance/                        | Outcome of differential abundance analysis for each S and mode.
-- figures/                                       | Figures (most of the used in the paper).
-- merged_counts/                                 | Stats related to the normalization process.
-- profiling/                                     | Output of profiling for each mode, sample and profiler. The output is provided as TAXPASTA output for genus and species, plus additional txt files from the profiler.
-- summary/                                       | Count tables for each sample, mode, S and normalization combination.

src/                                              | Source files.
|                                               
-- version_2/                                     | This is the latest version.
|    |
|    -- install/                                  | Files to prepare the environment and run the profiling.
|    |   |
|    |   -- build_pys/                            | Python scripts to create the necessary file for the generation of profiler DBs.
|    |   -- run_profilers/                        | Functions used to run profilers.
|    |   -- X_*.sh                                | bash scripts to run all the profilers.
|    |   -- list_vars.sh                          | List of variables used for the rest of .sh files in the folder.
|    |   -- sample_processing_pipeline.sh         | This is the general sh file to run X_*.sh files from 0 to 2 (3 is excluded).
|    -- sh_funcs/                                 | Derivative bash functions.
|    -- table_artificial_taxid.csv                | File with all information to create the in silico sample.   
|    -- list_vars.py                              | List of variables used for the rest of .ipynb files.
|    -- X_*.ipynb                                 | Notebook files to run the analysis of the profiling.

How to Run the Profiling

1. Preparation

1. Clone the Repository:

   git clone https://github.com/NanoNeuro/EV_taxprofiling.git
   cd EV_taxprofiling/src/version_2/install

2. Set Up the Environment: Install necessary dependencies for the project. You can create a conda environment using the provided environment.yml file (if available) or install the necessary packages manually.

   conda env create -f environment.yml  # Replace with your environment setup command if different
   conda activate EV_taxprofiling

3. Download the files: You can download the files used for this experiment from GEO database (GSE255317). You should download the fastq files into the data/. Maybe you need to rename them; you can do it as in data/samples_rnaseq.csv. Also feel free to adapt the code for your samples!!

2. Create the Profiler Database

To create the profiler database, navigate to the build_pys directory and run the relevant Python scripts. This step prepares the required files for database generation.

You can adapt your scripts to include the profilers you deem necessary.

1. Navigate to the build_pys Directory:

   cd build_pys

2. Run the Scripts: Execute the Python scripts to generate the necessary database files. Replace <script_name> with the actual script names provided in the folder.

   python <script_name>.py

3. Stuff to consider

• The species are downloaded using this command: ncbi-genome-download -F protein-fasta,fasta -p ${CPUS} -r 10 -P -l complete,chromosome -o ${BASEDIR_PROFILER_DB}/GENERAL/archaea --flat-output -m ${BASEDIR_PROFILER_DB}/GENERAL/archaea/table_archaea.txt archaea. This will download ~50k bateria and ~15k virus, which may be too much. You can change the -l parameter to include fewer species.
• The process to create certain DBs is resource extensive! Maybe you need a powerful computer or an HPC to run this part. Unfortunately, the databases are so big that I cannot save them anywhere.

3. Run the Profilers

1. Set Up Variables: Edit the list_vars.sh file to set any necessary variables for your analysis. Open it in a text editor and customize it as needed.

   nano list_vars.sh

   Save and exit.

2. Run the Bash Scripts: Execute the .sh scripts in the run_profilers directory to initiate the profiling process.

   bash X_<script_name>.sh

   Ensure you have the necessary permissions to run the scripts.

   chmod +x X_<script_name>.sh

3. Monitor Progress: Profiling may take some time depending on the size of your dataset. Check the output logs for progress and troubleshoot any errors.

How to Run the Jupyter Notebooks

1. Open and Execute Notebooks

1. Select the desired notebook (e.g., X_<notebook_name>.ipynb).

2. Follow the instructions within the notebook cells. Make sure to execute them in sequence.

3. Edit the list_vars.py file as necessary to set variables for analysis, similar to the Bash workflow.

4. How to replicate the analysis All required intermediate files are downloadable from Zenodo 10.5281/zenodo.14887264. You can download them and decompress them following the directory structure indicated at the top.

Cite us

 A proposed workflow to analyze bacterial transcripts in RNAseq from blood extracellular vesicles of people with Multiple Sclerosis
Alex M. Ascensión, Miriam Gorostidi-Aicua, Ane Otaegui-Chivite, Ainhoa Alberro, Rocio del Carmen Bravo-Miana, Tamara Castillo-Trivino, Laura Moles, David Otaegui
Frontiers in Microbiology; doi: https://doi.org/10.3389/fmicb.2025.1486661