DeepVariant¶

DeepVariant is an analysis pipeline that uses a deep neural network to call genetic variants from next-generation DNA sequencing data. DeepVariant relies on Nucleus, a library of Python and C++ code for reading and writing data in common genomics file formats (like SAM and VCF) designed for painless integration with the TensorFlow machine learning framework.

CPU 版本的 Singularity DeepVariant¶

CPU版安装¶

申请计算节点，然后制作 singularity 镜像

$ srun -p cpu -N 1 --exclusive --pty /bin/bash
$ singularity build deepvariant.simg docker://google/deepvariant

用SLURM脚本提交CPU版DeepVariant作业¶

使用 CPU 版本的 singluarity 镜像的 slurm.sh 如下：

#!/bin/bash

#SBATCH -J DeepVariant
#SBATCH -p small
#SBATCH -n 1
#SBATCH --ntasks-per-node=1
#SBATCH -o %j.out
#SBATCH -e %j.err

ulimit -s unlimited
ulimit -l unlimited

IMAGE_PATH=/安装路径/deepvariant.simg

singularity run $IMAGE_PATH /opt/deepvariant/bin/make_examples

并使用如下指令提交：

$ sbatch slurm.sh

交互式提交CPU版DeepVariant作业¶

srun -p cpu -N 1 --exclusive --pty /bin/bash
export IMAGE_PATH=/安装路径/deepvariant.simg
singularity run $IMAGE_PATH /opt/deepvariant/bin/make_examples

GPU 版本的 Singularity DeepVariant¶

GPU版安装¶

申请计算节点，然后制作 singularity 镜像

$ srun -p cpu -N 1 --exclusive --pty /bin/bash
$ singularity build deepvariant.gpu.simg docker://google/deepvariant:0.10.0-gpu

用SLURM脚本提交GPU版作业¶

使用GPU版本的 singluarity 镜像的 slurm.sh 如下：

#!/bin/bash

#SBATCH -J DeepVariant
#SBATCH -p dgx2
#SBATCH -N 1
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=6
#SBATCH --gres=gpu:1
#SBATCH --mem=MaxMemPerNode
#SBATCH -o %j.out
#SBATCH -e %j.err

ulimit -s unlimited
ulimit -l unlimited

IMAGE_PATH=/安装路径/deepvariant.gpu.simg

singularity run $IMAGE_PATH /opt/deepvariant/bin/make_examples

并使用如下指令提交：

$ sbatch slurm.sh

交互式提交GPU版deepvarant作业¶

srun --ntasks-per-node=1 -p dgx2 --gres=gpu:1 -N 1 --pty /bin/bash
export IMAGE_PATH=/安装路径/deepvariant.gpu.simg
singularity run $IMAGE_PATH /opt/deepvariant/bin/make_examples

参考资料¶

DeepVariant官网 https://github.com/google/deepvariant

最后更新: 2024 年 11 月 19 日