MiScan 1.0.0

Maxout-inferred SNV-based cancer prediction model

MiScan

Maxout-inferred SNV-based cancer prediction model | Apache Software License

Tutorial

Dependency data download

  To predict breast cancer risk with MiScan, Users firstly needs to do:

• VCF Files

  We recommend using whole exon sequencing data to obtain individual variation information, but the results of whole genome sequencing and full-length RNA-seq data are also compatible with the model. Please download test VCF files from FTP

• Maxout model weight

  Users also need a MiScan model weight to perform prediction, well-trained model weight can be downloaded from here

installation

• install through pip

  for linux or Mac OS user, MiScan can be installed easily using pip

pip install MiScan -i https://pypi.python.org/pypi

• install through docker

  for windows user, we provide a Docker version for convenient.

docker pull jefferyustc/miscan_command_line:v0.2.1
docker run --name miscan_cli_test -it -v /path/to/data:/path/in/docker 9fd

  More over, Docker Fileis available in the project directory, build it youself if you'd like to.

usage-commandline

Suppose your VCF file and weight are placed in the $dir directory.

MiScan --vcf $dir/SRR5447191.combined.filtered.vcf -o outputs --weight $dir/_MiScan_weights.hdf5

or with below command:

python -m MiScan --vcf $dir/SRR5447191.combined.filtered.vcf -o outputs --weight $dir/_MiScan_weights.hdf5

if with docker, the path of VCF file or weight path shoule be path in Docker environment:

MiScan -o test_outputs --vcf $dir_in_docker/SRR5447191.combined.filtered.vcf --weight $dir_in_docker/_MiScan_weights.hdf5

usage-script

from MiScan import miscan_main

miscan_main(
    outDir='./outputs',
    inVcf='/Users/jeffery/Downloads/SRR5447191.combined.filtered.vcf',
    model_weight='/Users/jeffery/workspace/projects/outputs/_MiScan_weights.hdf5'
)