AlphaFold2 运行验证

ice-melt

运行 alphafold

下载数据库

bfd|(6=[a3m|cs219|hhm].[ffdata|ffindex])#~1.8T(1.7T?)
mgnify|(1=[mgy_clusters_2018_12.fa])#~64G
params|(16=[5 CASP14 models|5 pTM models|5 AlphaFold-Multimer models|LICENSE])#~5.3G
pdb70|(9=pdb70_[a3m|cs219|hhm].[ffdata|ffindex],pdb70_clu.tsv,pdb_filter.dat,md5sum)#~56G(56G)
pdb_mmcif|(TODO: 数据不完整x=[mmcif_files/about 180,000 .cif files,obsolete.dat])#~G(206G)
pdb_seqres|(1=[pdb_seqres.txt])#~217M(0.2G)
uniclust30(13=[uniclust30_2018_08/])#~87G(86?)
uniprot|(1=[uniprot.fasta])#~104G(98.3?)
uniref90|(1=[uniref90.fasta])#~63G(58?)
small_bfd|(1=[bfd-first_non_consensus_sequences.fasta])#~17G(17G)

# 使用链接下载需要的额外处理
# bfd *****************************************************
BASENAME=bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt.tar.gz
tar --extract --verbose --file="${BASENAME}" \
  --directory=./
rm "${BASENAME}"
# mgnify *****************************************************
gunzip mgy_clusters_2018_12.fa.gz
# params *****************************************************
tar --extract --verbose --file=alphafold_params_2022-01-19.tar \
  --directory=./ --preserve-permissions
rm alphafold_params_2022-01-19.tar
# pdb70 ****************************************
tar --extract --verbose --file=pdb70_from_mmcif_200401.tar.gz \
  --directory=./
rm pdb70_from_mmcif_200401.tar.gz
# pdb_mmcif ****************************************
find "raw/" -type f -iname "*.gz" -exec gunzip {} +
mkdir mmcif_files
find raw/ -type d -empty -delete  # Delete empty directories.
for subdir in raw/*; do
  mv "${subdir}/"*.cif mmcif_files
done
find raw -type d -empty -delete
# download obsolete.dat
# pdb_seqres ****************************************
# 只有一个txt文本文件,无需处理
# uniclust30 ****************************************
tar --extract --verbose --file=uniclust30_2018_08_hhsuite.tar.gz \
  --directory=./
rm uniclust30_2018_08_hhsuite.tar.gz
# uniprot ****************************************
gunzip uniprot_trembl.fasta.gz
gunzip uniprot_sprot.fasta.gz
cat uniprot_sprot.fasta >> uniprot_trembl.fasta
mv uniprot_trembl.fasta uniprot.fasta
rm uniprot_sprot.fasta
# uniref90 ****************************************
gunzip uniref90.fasta.gz

准备运行环境

安装 nvidia 容器工具包支持GPU

安装指导

curl https://get.docker.com | sh \
  && sudo systemctl --now enable docker

distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
   && curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
   && curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

curl -s -L https://nvidia.github.io/nvidia-container-runtime/experimental/$distribution/nvidia-container-runtime.list | sudo tee /etc/apt/sources.list.d/nvidia-container-runtime.list

sudo apt-get update
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker

sudo docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi
# 展示nvidia gpu信息 就OK

设置非root用户运行docker

Docker as a non-root user

# cat /etc/group | grep docker
# 
sudo groupadd docker
sudo usermod -aG docker $USER
# 重启系统,或执行以下命令激活组修改
newgrp docker 
# 验证
docker run hello-world

# 出现该错误，修改 ~/.docker 的权限
WARNING: Error loading config file: /home/user/.docker/config.json -
stat /home/user/.docker/config.json: permission denied

sudo chown "$USER":"$USER" /home/"$USER"/.docker -R
sudo chmod g+rwx "$HOME/.docker" -R

检查AlphaFold是否可以使用GPU

docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi

构建docker镜像

docker build -f docker/Dockerfile -t alphafold .

修改Dockerfile 以解决以下问题

apt-get update失败

RUN echo deb https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 / | tee /etc/apt/sources.list.d/cuda.list
# apt-get update 命令之前添加以上命令

修改hh-suite仓库代码路径

#RUN git clone --branch v3.3.0 https://github.com/soedinglab/hh-suite.git /tmp/hh-suite \
RUN git clone --branch v3.3.0 https://github.91chi.fun/https://github.com/soedinglab/hh-suite.git /tmp/hh-suite \

pip 下载太慢

# 将这几个命令分开,成功的步骤下次build时可以使用cache
# tensorflow 太大,注释 requirements.txt 中对 tensorflow 的依赖
# tensorflow-cpu==2.5.0 改为 tensorflow==2.5.0,使同时支持cpu和gpu
RUN pip3 install --upgrade pip
RUN pip3 install -r /app/alphafold/requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simplie
RUN pip3 install tensorflow==2.5.0 -i https://pypi.tuna.tsinghua.edu.cn/simplie
RUN pip3 install --upgrade jax==0.2.14 jaxlib==0.1.69+cuda${CUDA/./} -f \
      https://storage.googleapis.com/jax-releases/jax_releases.html

其它准备

# docker/run_docker.py
flags.DEFINE_string(
    'output_dir', '/tmp/alphafold',
    'Path to a directory that will store the results.')
# 需要修改路径,修改这里
# 'output_dir', '/ceph_disk2/MetaDatabase/AlphaFold_DB/tmp',
mkdir -p /ceph_disk2/MetaDatabase/AlphaFold_DB/tmp
cd /ceph_disk2/MetaDatabase/AlphaFold_DB
chmod 770 tmp

修改镜像名

# docker/run_docker.py
flags.DEFINE_string(
    'docker_image_name', 'alphafold', 'Name of the AlphaFold Docker image.')
# 修改为
'docker_image_name', 'anneng01:8090/library/alphafold'

准备python环境

conda create -n af2 -y python=3.7
conda activate af2
pip3 install -r docker/requirements.txt

运行验证

1. 准备数据

CASP14 target list
找到T1050,复制该序列到文件 T1050.fasta

2. 运行

python3 docker/run_docker.py \
  --fasta_paths=data/T1050.fasta \
  --max_template_date=2020-05-14 \
  --data_dir=/ceph_disk2/MetaDatabase/AlphaFold_DB/genetic_databases

3. 监控运行状态,查看运行结果

watch -n 1 nvidia-smi

https://www.ncbi.nlm.nih.gov/Structure/icn3d/full.html
File>Open File>PDB File(appendable)> 选择文件> Append

ice-melt

pdb_mmcif 下载特别慢

# 参考网站 https://www.wwpdb.org/ftp/pdb-ftp-sites
# 找到合适的链接下载,比如:
rsync -rlpt -v -z --delete \
rsync.ebi.ac.uk::pub/databases/pdb/data/structures/divided/mmCIF/ \
./mmCIF

# 脚本中给出的信息:
#echo "  * rsync.ebi.ac.uk::pub/databases/pdb/data/structures/divided/mmCIF/ (Europe)"
#echo "  * ftp.pdbj.org::ftp_data/structures/divided/mmCIF/ (Asia)"
#echo "or see https://www.wwpdb.org/ftp/pdb-ftp-sites for more download options."
#rsync --recursive --links --perms --times --compress --info=progress2 --delete --port=33444 \
#  rsync.rcsb.org::ftp_data/structures/divided/mmCIF/ \
#  "${RAW_DIR}"