A demo server for RAWG is available here. Note that due to resource constraint, this is only meant to demonstrate the front-end capability of RAWG, hence analysis workflows will not run.
In RAWG's setup, web admin can provide pre-generated genome index files for analysis. Users hence have two choices when creating a new session 1) use provided genome files 2) upload own genome and annotation files.
In the example below, Preindexed genome files are selected for Homo Sapiens. After creating the session, user will be directed to the SESSION INFO tab which shows user the overview infomation and an identifier so that user can retrieve this session from the home page.
After creating the session, user can add sample conditions (contrast) and specify number of replicates from the SAMPLE tab. An arbitary number of conditions can be added and RAWG will conduct pairwise differential analysis for all the possible contrast pairs.
RAWG is configure to take both single reads and paired-end reads although it's not a good idea to mix sample types in the same comparison. RAWG can take fastq files or compressed fastq files (.gz format).
After uploading all samples and specified the conditions, the user is then directed to define workflows. From the WORKFLOW tab, user can add an arbitary number of workflows.
In this example, three workflows were defined
- STAR - Stringtie - DESeq2
- STAR - Stringtie - edgeR
- HISAT2 - Stringtie - edgeR
This shows that RAWG is able to reused the common results (STAR - Stringtie) for different downstream analysis.
Finally, the defined session is submitted. A workflow schematic is generated and displayed in the SESSION INFO tab. This schematic tries to follow CWL viewer's style. At the bottem of the same tab, the schematic, cwl workfklow file and input file can be downloaded. The cwl files can be used for local execution as long as files paths are correctly resolved. Note that the workflow generation script runs every 15 s on the server so you may need to wait 15 s at most.
KNOWN ISSUE: the workflow schematic will not display automatically and the user needs to refresh the webpage for the diagram to load correctly.
We have implemented some minimal visualisation functionalities, volcano plot and bar plot. The visualisation page is accessible from the bottem of the SESSION INFO tab
The volcano plot allows user to quickly pick out the significant genes and more than one pair of contrast can be shown on the same diagram. Users can also adjust the p-value and logfold change threshold interactively. Mouse hovering over a dot on the plot shows the gene accession and relevant data at the bottom of the plot.
The bar plot is a simple diagram that tells user the number of significant genes found based on user-selected parameters.
To try out the visualisation functions yourself, you can use identifier 4e4dd873-f4a8-45a6-8a74-7b67173568b4 to input at the homepage and access this session. Or you can click this link directly.
To produce the plot, you need to first select data from the selection menu on the left then click the Volcano Plot or Bar Plot tab heading. Note that some complex data set can take ~ 1 min to plot.
The demo is hosted on a DigitalOcean droplet with minimum comuptational resource hence it is only meant to be a demonstration for the front-end interface. This droplet uses Ubuntu 18.04 as base image.
Here is the history shows commands used to set up the demo server
root@rawg:~# history
1 git clone --recurse-submodules -j3 https://github.com/rawgene/rawg
2 apt update
3 apt install graphviz python3-pip
4 cd rawg/webportal/webportal/
5 cp settings.py local_settings.py
6 vim local_settings.py
7 cd ../..
8 pip3 install -r requirements.txt
9 cd webportal/
10 python3 manage.py makemigrations
11 python3 manage.py migrate
12 nohup python3 manage.py runserver 0.0.0.0:80 &