Tuturial for running Alphafold on Tanglab.pku.edu.cn

---

Query

alphafold vs standard structure

---

Example

绿色tanglab.pku.edu.cn，蓝色北极星

Dimer

蓝色部分是晶体结构

---

---

Start Alphafold-Multitimer

💡 ！！！注意！文件名应使用英文且不含空格！！

Prepare .fasta

Folding a homomer

>sequence_1
<SEQUENCE>
>sequence_2
<SEQUENCE>
>sequence_3
<SEQUENCE>

Folding a heteromer

>sequence_1
<SEQUENCE A>
>sequence_2
<SEQUENCE A>
>sequence_3
<SEQUENCE B>
>sequence_4
<SEQUENCE B>
>sequence_5
<SEQUENCE B>

Start Fold!

our alphafold2 is installed in /smb/jiangyd/alphafold

权限: 777,请勿删除文件！

# log in tanglab.pku.edu.cn 
# bash shell

# activate alphafold python environment
module load conda/anaconda3-2021-05
source activate
conda activate alphafold

# cd root dir
## !! run_alphafold.sh must be run in the root directory!！
cd /smb/jiangyd/alphafold-multitimer/alphafold/

# code to start alphafold! 
## for monomer(Single Protein)
bash /smb/jiangyd/alphafold-multitimer/alphafold/run_alphafold.sh \
-d /smb/jiangyd/alphafold/alphafold_data/ \
-o /your_output_dir \
-m monomer \
-f /smb/jiangyd/alphafold/your_fasta_file \
-t 2021-11-20 \


## for multimer
bash /smb/jiangyd/alphafold-multitimer/alphafold/run_alphafold.sh \
-d /smb/jiangyd/alphafold/alphafold_data/ \
-o /your_output_dir \
-m multimer \
-f /smb/jiangyd/alphafold/your_fasta_file \
-t 2021-11-20 \


# Test code
bash /smb/jiangyd/alphafold-multitimer/alphafold/run_alphafold.sh -d /smb/jiangyd/alphafold/alphafold_data/ -o /home/jiangyd/test/ -m monomer -f /smb/jiangyd/alphafold/example/query.fasta -t 2021-09-23 -g False

or you can make a script(such as run_alphafold_yida.sh) in your love directory

# Auto run alphafold bash script
module load conda/anaconda3-2021-05
source activate
conda activate alphafold
cd /smb/jiangyd/alphafold-multitimer/alphafold/

# the code below can be edited
nohup bash /smb/jiangyd/alphafold-multitimer/alphafold/run_alphafold.sh \
-d /smb/jiangyd/alphafold/alphafold_data/ \
-o /home/jiangyd/test/ \
-m monomer \
-f /smb/jiangyd/alphafold/example/query.fasta \
-t 2021-12-07 &> nohup_20211207.out&  
echo$(echo -e "\n")
## the output will be saved in nohup file in /smb/jiangyd/alphafold-multitimer/alphafold/
## the code will be run in background

cd -
## back to the directory before

Model

对于上面的 -m 选项，有以下的备选方法：

1. You can control which AlphaFold model to run by adding the -model_preset= flag. We provide the following models:
   • monomer: This is the original model used at CASP14 with no ensembling.
   • monomer_casp14: This is the original model used at CASP14 with num_ensemble=8, matching our CASP14 configuration. This is largely provided for reproducibility as it is 8x more computationally expensive for limited accuracy gain (+0.1 average GDT gain on CASP14 domains).
   • monomer_ptm: This is the original CASP14 model fine tuned with the pTM head, providing a pairwise confidence measure. It is slightly less accurate than the normal monomer model.
   • multimer: This is the AlphaFold-Multimer model. To use this model, provide a multi-sequence FASTA file. In addition, the UniProt database should have been downloaded.

run_alphafold.sh的具体参数

Usage: ./run_alphafold_v21.sh <OPTIONS>
Required Parameters:
-d <data_dir>         Path to directory of supporting data
-o <output_dir>       Path to a directory that will store the results.
-f <fasta_path>       Path to a FASTA file containing sequence. If a FASTA file contains multiple sequences, then it will be folded as a multimer
-t <max_template_date> Maximum template release date to consider (ISO-8601 format - i.e. YYYY-MM-DD). Important if folding historical test sets
Optional Parameters:
-g <use_gpu>          Enable NVIDIA runtime to run with GPUs (default: true)
-n <openmm_threads>   OpenMM threads (default: all available cores)
-a <gpu_devices>      Comma separated list of devices to pass to 'CUDA_VISIBLE_DEVICES' (default: 0)
-m <model_preset>     Choose preset model configuration - the monomer model, the monomer model with extra ensembling, monomer model with pTM head, or multimer model (default: 'monomer')
-c <db_preset>        Choose preset MSA database configuration - smaller genetic database config (reduced_dbs) or full genetic database config (full_dbs) (default: 'full_dbs')
-p <use_precomputed_msas> Whether to read MSAs that have been written to disk. WARNING: This will not check if the sequence, database or configuration have changed (default: 'false')
-l <is_prokaryote>    Optional for multimer system, not used by the single chain system. A boolean specifying true where the target complex is from a prokaryote, and false where it is not, or where the origin is unknown. This value determine the pairing method for the MSA (default: 'None')
-b <benchmark>        Run multiple JAX model evaluations to obtain a timing that excludes the compilation time, which should be more indicative of the time required for inferencing many proteins (default: 'false')

• optionally set the -is_prokaryote_list flag with booleans that determine whether all input sequences in the given fasta file are prokaryotic. If that is not the case or the origin is unknown, set to false for that fasta.

Scripts

How to run a number of fasta files at once?

# make sure you are in right dir
cd /smb/jiangyd/alphafold

# make a folder to put your fasta files
mkdir example
scp ……（略）

# write a script to start alphafold!
# eg. run_ubtest.sh(in /smb/jiangyd/alphafold !)
vim run_ubtest.sh
## 以下部分需要你在.sh脚本文件里手动编辑

for fastafile in ubtest/*
do

    echo "##############################"
    echo "##############################"
    echo "##############################"
    date +"%Y-%m-%d-%H-%M-%S"
    echo $fastafile
    echo "##############################"
    echo "##############################"
    echo "##############################"
		bash /smb/jiangyd/alphafold-multitimer/alphafold/run_alphafold.sh -d /smb/jiangyd/alphafold/alphafold_data/ -o /your_output_dir -m monomer -f /smb/jiangyd/alphafold/your_fasta_file -t 2021-09-23 -g False
done

#save scripts

后台运行alphafold并记录输出文档到nohup_20210924ubtest.out

nohup bash run_ubtest.sh &> nohup_20210924ubtest.out&  

Monitor

##(可选)
nohup vmstat 10 &> nohup_20210924ubtest_vmstat.out&

#如果你想看程序有没有在跑，输入top查看有无高占用进程即可

Reference

GitHub - deepmind/alphafold: Open source code for AlphaFold.

https://github.com/kalininalab/alphafold_non_docker

CASP14：谷歌DeepMind的AlphaFold 2到底取得了什么成就? 它对蛋白质折叠, 生物学和生物信息学意味着什么?

---