Interactive visualization of bacterial pan-genomes

Ensure you have npm installed:

brew install npm          # MacOS Homebrew or Linuxbrew
sudo apt-get install npm  # Debian/Ubuntu/Mint
sudo yum install npm      # Redhat/Centos/Fedora

Install the code and build:

git clone https://github.com/drpowell/FriPan
cd FriPan
npm install
make compile

Run the demo code locally using our Python-based server in server.sh:

./server.sh
firefox http:https://localhost:8030/pan.html

Or make it publicly accessible using your public_html if your server is already running Apache:

# will put in $HOME/public_html/fripan by default, type 'make help' to see options
make install
firefox http:https://localhost/~user/fripan/pan.html

After you've run a pan-genome analysis with roary, copy the generated gene_presence_absence.csv file to the FriPan root directory and name it some_strain.roary (eg ecoli.roary).

Go to http:https://localhost:8030/pan.html?some_strain (where some_strain is the name you chose for the file above).

Tells you how many isolates were loaded and how many gene clusters across those isolates, as well as links to error logs and software information.

This is not a phylogenetic tree. It is constructed from binary matrix of gene presence/absence for each pair of gene cluster and isolate. It dynamically changes depending on the selection in panel E/F.

This MDS plot is a dimensionality scaling plot (like PCA) which uses gene presence/absence to group isolates with similar accessory genomes. Note that core genome has no influence on this plot because those genes are present in all the isolates. This panel also dynamically updates to represent the current selection in panel E/F.

This graph plots the percentage of the signal present in each of the MDS dimensions. Ideally most of the signal will be in the first 2 dimensions which are shown in panel C1. You can click on any of the bars to change which dimensions are displayed in panel C1.

These options allow you to configure how the pan-genome, tree and MDS plot are displayed. You can re-order rows, annotate with colours etc.

This panel will always show a zoomed out version of the whole pan-genome. The x-axis shows the gene clusters - here we see we have just over 900 clusters, which matches the 931 denoted in Panel A. It is possible to select a region of this panel to zoom in on a smaller section of the pan-genome.

This panel provides a scrollable view of the pan genome itself. When zoomed in, each gene is visible as a block, with width approximaltely proportional to the gene length. Hovering over a gene will display information about that gene that was present in the input files.

Coming soon.

Coming soon.

An example set of input files with the stem test is provided:

1. test.proteinortho
2. test.descriptions
3. test.strains

This is the gene presence/absence matrix in TSV format. Each row is a gene ortholog cluster, and each column in a strain. Each cell in the matrix is gene ID, or * if none. Paralogs are CSV within the cell. The first 3 columns are unused, but you must use the exact names as below.

# Species   Genes   Alg.-Conn.   USA300    TW20      JKD6159
3           3       1            USA_001   TW20_001  JKD_001  
3           4       1            USA_002   TW20_002  JKD_002,JKD_004
2           2       1            USA_003   *         JKD_003 
1           1       1            USA_004   *         *

This maps gene IDs from the strain columns, in 2-column TSV format.

USA_001	      DNA replication protein
USA_002       hypothetical protein
USA_003       gyrase A
USA_004       alcohol dehydrogenase (EC:1.1.1.1)
TW20_001      DNA replication protein
TW20_002      unknown protein
JKD_001       DNA replication protein, dnaA
JKD_002       hypothetical protein
JKD_003       gyrase
JKD_004       hypothetical protein

This a multi-column TSV format. The first ID column links it with the strains in the other two files. The remainign columns can be used for colouring and ordering within the application.

ID        ST     Phenotype     Country    Colour
USA300    239    resistant     US         blue
TW20        2    suspectible   UK         green
JKD6159   239    resistant     AU         red

The included example input files all start with the stem/prefix test. You can add as many pan-genomes to the fripan folder as you like, just give each of them a different stem, say mypop. Then you just append ?mypop to the URL, so it looks like http:https://example.com/~user/fripan/pan.html?mypop.

To simplify this, just add each stem to the file called pan.index. This will allow them to be selected via the Index menu item within the application.

While developing code, it is useful to enable coffee in "watch" mode and with source maps. Run the following:

make debug

Report feedback, suggestions and bugs on the Issues page

MIT

• Github: https://github.com/drpowell/FriPan
• Website: http:https://drpowell.github.io/FriPan/

• David Powell
• Torsten Seemann