mouse-pergola-reproduce.nf

This repository contains the software, scripts and data to reproduce the results corresponding to the high-fat diet mice experiment of the Pergola paper. The original source of this data set can be found on this paper

If you have not install yet docker and nextflow, follow this intructions

Clone the repository

git clone --recursive https://github.com/cbcrg/mouse-pergola-reproduce.git
cd mouse-pergola-reproduce

Data

Data is publicly available in Zenodo as a compressed tarball .

Data can be downloaded and uncompressed using the following command:

mkdir data
wget -O- https://zenodo.org/record/1300200/files/mouse_dataset.tar.gz | tar xz -C data

Pull docker image

Pull the Docker image use for processing data with Pergola (Pergola and its dependencies installed)

docker pull pergola/pergola-reproduce@sha256:02bf3e701175104a488f40761b856efa1f97e2f2f82af8adae63b24ac2517326

Run nextflow pipeline

Once data is downloaded, it is possible to reproduce all the results using this command:

NXF_VER=0.30.2 nextflow run mouse-pergola-reproduce.nf \
  --recordings='data/mouse_recordings/' \
  --mappings='data/mappings/b2p.txt' \
  --mappings_bed='data/mappings/bed2pergola.txt' \
  --phases='data/phases/exp_phases.csv' \
  --mappings_phase='data/mappings/f2g.txt' \
  --exp_info='data/mappings/exp_info.txt' \
  --tbl_chromHMM="data/chromHMM_files/cellmarkfiletable" \
  --n_bins_HMM=5 \
  --n_states_HMM=4 \
  --image_format='png' \
  -with-docker

Results

The previous command generates a results folder that contains the results of the analysis shown in the paper:

• A figure created using Gviz displaying the mouse feeding behavior and mouse accumulated food intakes within 30 minutes time-windows displayed as a heatmap.
• A figure created using Sushi rendering the same layout as in the previous figure.
• A folder containing all the necessary files to render the raw feeding behavior and the the accumulated food intakes using IGV. Data is separated in folders corresponding to each mouse group.
• A heatmap comparing the feeding behavior of high-fat mice group with their controls (fold change).
• A folder named feeding_activity_profiles containing the actograms generated using deepTools.
• All files regarding the HMM modeling of the data using chromHMM can be found on chromHMM folder.
• A folder named files which contains the complete set of files needed to visualize the data using Shiny-pergola as explained below.

Online Shiny-Pergola visualization

From the folder where you have run the pipeline, you can visualize the results using Shiny-Pergola introducing the commands described below.

shiny-pergola config file

To set the membership of each of the files inside the folder with the same name, a config file is needed. The repository includes a file assigning files to each of the groups (Control, HF) You can also check the file here

Downloading and running the shiny-pergola image

Pull the docker image containing the version of shiny-pergola web application used for render the data visualization:

docker pull pergola/shiny-pergola@sha256:e4c470b3916ce6ce0298b231d3a8d18bf72535beb04251c13439351c90797431

With docker running, launch the image:

docker run --rm -p 3600:80 -v "$(pwd)":/pergola_data pergola/shiny-pergola@sha256:e4c470b3916ce6ce0298b231d3a8d18bf72535beb04251c13439351c90797431 &

Note: "$(pwd)" can be substitute by your absolute path to the folder where the mouse-pergola-reproduce.nf has been run.

Note: Paper contain several figures related to mouse dataset analysis. If you want to get exactly one of the figure just select it by setting the figure on id.txt file. For instance if you want to reproduce Supplementary figure 4a you just have to type the following command before running Docker shiny-pergola image.

echo "hf_1" > id.txt

The codes for each figure are:

Figure	Code
Fig. 4a	hf_1
Fig. 4b	hf_2
Supp. Fig. S2	hf_3

Go to your web browser and type in your address bar the following ip address: http://0.0.0.0:3600

Note: In newer Docker versions by default the IP address used is the localhost 0.0.0.0. As this IP might be used by other services in the default port (80), we especified the port used in our docker command by -p 3600:80. Old docker versions might used a different IP address that you can get using the following command (usually this one http://192.168.99.100). You may then enter this IP followed by the same port in the address bar of your browser e.g. http://192.168.99.100:3600.

docker-machine ip default 

Et voila, the running container will load the shiny app in your browser.