PAVOOC

Prediction and visualization of on- and off-targets for CRISPR

CONTRIBUTIONS welcome: The software should be easy to set up and work with on your local machine. If you have feature requests, file an issue (along with ideas on how to realize it).

The platform is based on hg19. hg38 is used only for the model generation to be in sync with the azimuth dataset.

Install

python3 with pip, mongodb and mongodb-tools (mongorestore) and npm are required.

pip install -r requirements.txt
cd webapp && npm install

Alternatively you can run everything in docker.

Run it

python -m pavooc.pipeline

builds up all required data

python -m pavooc.server.main

and cd webapp && npm start

bring up a local version of the webapp

Note: training of the PAVOOC DNN requires installation of (cuda) pytorch (http://download.pytorch.org/whl/cpu/torch-0.3.1-cp36-cp36m-linux_x86_64.whl) Note 2: By default a dump of the database is used to initialize all required data. To compute everything by scratch (probably takes several days), use the ONLY_INIT=0 environment variable (see pipeline.py and docker-compose.yml)

Docker

Alternatively there is a docker-environment which should get you started with

docker-compose pull
docker-compose up

Conventions

this is all hg19/GRCh37!
In the DB 'exons' right now are only the exons for the canonical transcript.. this also means, that in knockout and edit experiments we only consider the canonical transcripts
canonical exons contain all exons of the canonical transcript MINUS UTR regions
All sequences are saved in gene-relative direction ('-'-strands are reverse-complemented).
Ensembl gene IDs are used
Exon are saved with paddings such that a sgRNA search can find all guides that would cut inside the exon. Forward strand: (16padding)(exon)(6padding), Backward strand: (6padding)(exon)(16padding)

Datasources

What follows is a list of used datasets

Reference genome GRCh37

http://hgdownload.soe.ucsc.edu/goldenPath/hg19/chromosomes/

UCSC Pfam

Preprocessed Pfam data which maps protein domains to ucsc coordinates.

ftp://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/ucscGenePfam.{sql,txt.gz}

Gene annotations (exons included)

http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/refGene.txt.gz bin smallint(5) unsigned NOT NULL, name varchar(255) NOT NULL, chrom varchar(255) NOT NULL, strand char(1) NOT NULL, txStart int(10) unsigned NOT NULL, txEnd int(10) unsigned NOT NULL, cdsStart int(10) unsigned NOT NULL, cdsEnd int(10) unsigned NOT NULL, exonCount int(10) unsigned NOT NULL, exonStarts longblob NOT NULL, exonEnds longblob NOT NULL, score int(11) default NULL, name2 varchar(255) NOT NULL, cdsStartStat enum('none','unk','incmpl','cmpl') NOT NULL, cdsEndStat enum('none','unk','incmpl','cmpl') NOT NULL, exonFrames longblob NOT NULL,

Protein data

http://www.uniprot.org/downloads

Cancer cell line per-gene expression data

https://data.broadinstitute.org/ccle_legacy_data/dna_copy_number/CCLE_copynumber_byGene_2013-12-03.txt.gz

Deployment

Docker images

In docker-hub there are automated builds for pavooc and the configured nginx server. Default tag is latest and image names are moritzs/pavooc and moritzs/pavooc-nginx This speeds up deployment

docker-compose pull
docker-compose build # for nginx
docker-compose up -d

Manual builds

It is possible to manually build and push the images using

docker build -t pavooc .
docker tag pavooc moritzs/pavooc:latest
docker push moritzs/pavooc:latest
cd nginx
docker build -t pavooc-nginx .
docker tag pavooc moritzs/pavooc-nginx:latest
docker push moritzs/pavooc-nginx:latest

AWS docker machine

To create a EC2 instance with docker run

docker-machine create --driver=amazonec2 --amazonec2-instance-type t2.large --amazonec2-region eu-central-1 --amazonec2-root-size=250 machine-name

To connect to it run eval $(docker-machine env machine-name)

then simply run docker-compose up

Future work

adding necessary genome sequence tracks

For example lnc RNAs could be helpful for many..

exon expressions

GUIDES (http://guides.sanjanalab.org/) offers a nice way of considering exon expressions. It would be nice and easy to add an expression level to guides in the table in the GeneViewer table.

improve repository

Tests are not running a.t.m.. Also it would be nice to have automatic PEP8 and TypeScript sanity checks.

Model features

Adding chromatin accessibility could be a very useful feature..

Favicon

The website still has the stock react favicon. What a shame!

moritzschaefer / pavooc

readme