Difference between revisions of "AlphaFold-Sapelo2"

From Research Computing Center Wiki
Jump to navigation Jump to search
 
(39 intermediate revisions by the same user not shown)
Line 8: Line 8:
  
 
=== Version ===
 
=== Version ===
−
2.0.0, 2.0.1
+
2.2.4, 2.3.1, 2.3.4
  
 
=== Author / Distributor ===
 
=== Author / Distributor ===
Line 21: Line 21:
 
For more information on Environment Modules on Sapelo2 please see the [[Lmod]] page.
 
For more information on Environment Modules on Sapelo2 please see the [[Lmod]] page.
  
−
*Version 2.0.0, installed as a conda environment in /apps/gb/AlphaFold/2.0.0/
+
 
 +
*'''Version 2.2.4'''
 +
 
 +
This version is installed as a singularity container pulled from https://hub.docker.com/r/catgumag/alphafold:
 +
<pre class="gscript">
 +
/apps/singularity-images/alphafold_2.2.4.sif
 +
</pre>
 +
You can view the documentation for this version of AlphaFold with the following command, on an interactive node:
 +
<pre class="gcommand">
 +
singularity exec /apps/singularity-images/alphafold_2.2.4.sif python /app/alphafold/run_alphafold.py --helpfull
 +
</pre>
 +
This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).
 +
 
 +
The database files are installed in /apps/db/AlphaFold/2.2.4. To use these in the singularity container, please add the option <code>-B /apps/db/AlphaFold </code> to the singularity exec command, as shown in the sample job submission scripts below. The <code> --nv </code> option needs to be added to enable AlphaFold to run on the GPU. This singularity container also requires the option <code>--use_gpu_relax</code> to be added. 
 +
 
 +
 
 +
*'''Version 2.3.1 (on Intel nodes only)'''
 +
 
 +
This version is installed as a singularity container pulled from https://hub.docker.com/r/catgumag/alphafold:
 +
<pre class="gscript">
 +
/apps/singularity-images/alphafold_2.3.1_cuda112.sif
 +
</pre>
 +
You can view the documentation for this version of AlphaFold with the following command, on an interactive node:
 +
<pre class="gcommand">
 +
singularity exec /apps/singularity-images/alphafold_2.3.1_cuda112.sif python /app/alphafold/run_alphafold.py --helpfull
 +
</pre>
 +
This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).
 +
 
 +
The database files are installed in /apps/db/AlphaFold/2.3.1. To use these in the singularity container, please add the option <code>-B /apps/db/AlphaFold </code> to the singularity exec command, as shown in the sample job submission scripts below. The <code> --nv </code> option needs to be added to enable AlphaFold to run on the GPU. Please also add the option <code>-B /apps/eb/CUDAcore/11.2.1 </code>, so the singularity container can link to the CUDA libraries. This singularity container also requires the option <code>--use_gpu_relax</code> to be added. 
 +
 
 +
 
 +
*'''Version 2.3.1'''
 +
 
 +
Installed with EasyBuild in /apps/eb/AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0/
  
 
To use this version of AlphaFold, please first load the module with
 
To use this version of AlphaFold, please first load the module with
 
<pre class="gscript">
 
<pre class="gscript">
−
ml AlphaFold/2.0.0_conda
+
ml AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0
 
</pre>
 
</pre>
  
−
Once you load the module, an environmental variable called EBROOTALPHAFOLD is exported. It stores the AlphaFold installation path on the cluster, i.e., /apps/gb/AlphaFold/2.0.0. The bash script run_alphafold.sh in installed in EBROOTALPHAFOLD/alphafold, and the 2.2TB of database files are in /apps/db/AlphaFold (this is the directory that you need to use for the -d option of run_alphafold.sh).  
+
Once you load the module, an environmental variable called EBROOTALPHAFOLD is exported. It stores the AlphaFold installation path on the cluster, i.e., /apps/eb/AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0. The python script run_alphafold.py is installed in EBROOTALPHAFOLD/bin and a symbolic link called alphafold points to it and can be used to run the program. The 2.2TB of database files are in /apps/db/AlphaFold/2.3.1. You can export the environment variable ALPHAFOLD_DATA_DIR to set the location of the database files. For bash, use
 +
<pre class="gscript">
 +
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.1
 +
</pre>
 +
When you load the <code>AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0</code> module, this environment variable will be automatically set.
  
−
'''Note:''' This program does not work on the nodes with K20Xm GPU devices, because the CPUs on those nodes do not support AVX. If you run this program on the gpu_p partition, please request a K40 or a P100 GPU device.
+
'''Note:''' If you run this program on the gpu_p partition, please request a P100 or an A100 GPU device. This version requires a GPU device and it should work on the P100, V100, and A100 devices.  
  
−
*Version 2.0.1, installed with EasyBuild in /apps/eb/AlphaFold/2.0.1-fosscuda-2020b/
+
 
 +
 
 +
*'''Version 2.3.4'''
 +
 
 +
Installed with EasyBuild in /apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold/
  
 
To use this version of AlphaFold, please first load the module with
 
To use this version of AlphaFold, please first load the module with
 
<pre class="gscript">
 
<pre class="gscript">
−
ml AlphaFold/2.0.1-fosscuda-2020b
+
ml AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold
 
</pre>
 
</pre>
  
−
Once you load the module, an environmental variable called EBROOTALPHAFOLD is exported. It stores the AlphaFold installation path on the cluster, i.e., /apps/eb/AlphaFold/2.0.1-fosscuda-2020b. The python script run_alphafold.py is installed in EBROOTALPHAFOLD/bin and a symbolic link called alphafold points to it and can be used to run the program. The 2.2TB of database files are in /apps/db/AlphaFold. You can export the environment variable ALPHAFOLD_DATA_DIR to set the location of the database files. For bash, use
+
Once you load the module, an environmental variable called EBROOTALPHAFOLD is exported. It stores the AlphaFold installation path on the cluster, i.e., /apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold. The python script run_alphafold.py is installed in EBROOTALPHAFOLD/bin and a symbolic link called alphafold points to it and can be used to run the program. The 2.2TB of database files are in /apps/db/AlphaFold/2.3.4. You can export the environment variable ALPHAFOLD_DATA_DIR to set the location of the database files. For bash, use
 
<pre class="gscript">
 
<pre class="gscript">
−
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold
+
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.4
 
</pre>
 
</pre>
 +
When you load the <code>AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold</code> module, this environment variable will be automatically set.
 +
 +
'''Note:''' If you run this program on the gpu_p partition, please request a P100 or an A100 GPU device. This version requires a GPU device.
 +
 +
'''This version does not appear to be working now.'''
  
−
'''Note:''' This program does not work on the nodes with K20Xm GPU devices, because the CPUs on those nodes do not support AVX. If you run this program on the gpu_p partition, please request a K40 or a P100 GPU device.
 
  
 +
'''Sample Job Submission scripts'''
  
−
Sample job submission script (sub.sh) to run AlphaFold 2.0.0 using run_alphafold.sh in a batch job (without GPU):
+
Sample job submission script to run the singularity container for v. '''2.3.1''' for '''Monomer''' on a GPU:
  
 
<pre class="gscript">
 
<pre class="gscript">
 
#!/bin/bash
 
#!/bin/bash
 
#SBATCH --job-name=alphafoldjobname       
 
#SBATCH --job-name=alphafoldjobname       
−
#SBATCH --partition=batch           
+
#SBATCH --partition=gpu_p       
 
#SBATCH --ntasks=1                 
 
#SBATCH --ntasks=1                 
−
#SBATCH --cpus-per-task=4         
+
#SBATCH --cpus-per-task=4
−
#SBATCH --mem=20gb                   
+
#SBATCH --gres=gpu:P100:1
 +
#SBATCH --mem=50gb        
 +
#SBATCH --constraint=Intel           
 
#SBATCH --time=120:00:00           
 
#SBATCH --time=120:00:00           
 
#SBATCH --output=%x.%j.out     
 
#SBATCH --output=%x.%j.out     
 
#SBATCH --error=%x.%j.err           
 
#SBATCH --error=%x.%j.err           
−
#SBATCH --mail-user=username@uga.edu 
 
−
#SBATCH --mail-type=ALL 
 
  
 
cd $SLURM_SUBMIT_DIR
 
cd $SLURM_SUBMIT_DIR
  
−
ml AlphaFold/2.0.0_conda
+
ml purge
 +
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
 +
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0
  
−
bash $EBROOTALPHAFOLD/alphafold/run_alphafold.sh -d /apps/db/AlphaFold [options]
+
ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.1
−
</pre>  
+
 
 +
singularity exec -B /apps/db/AlphaFold -B /apps/eb/CUDAcore/11.2.1 \
 +
--nv /apps/singularity-images/alphafold_2.3.1_cuda112.sif python /app/alphafold/run_alphafold.py  \
 +
--use_gpu_relax \
 +
--data_dir=$ALPHAFOLD_DATA_DIR  \
 +
--uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta  \
 +
--mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa  \
 +
--bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt  \
 +
--uniref30_database_path=$ALPHAFOLD_DATA_DIR/uniref30/UniRef30_2021_03 \
 +
--pdb70_database_path=$ALPHAFOLD_DATA_DIR/pdb70/pdb70  \
 +
--template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files  \
 +
--obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
 +
--model_preset=monomer \
 +
--max_template_date=2022-10-01 \
 +
--db_preset=full_dbs \
 +
--output_dir=./output \
 +
--fasta_paths=./IL2Y.fasta
 +
</pre>
 +
 
 +
 
 +
Sample job submission script to run the singularity container for v. 2.3.1 for '''Multimer''' on a GPU:
  
−
An example of the required options to use are
 
 
<pre class="gscript">
 
<pre class="gscript">
−
bash $EBROOTALPHAFOLD/alphafold/run_alphafold.sh -d /apps/db/AlphaFold -o ./test/ -m model_1 -f ./query.fasta -t 2020-05-14
+
#!/bin/bash
 +
#SBATCH --job-name=alphafold
 +
#SBATCH --partition=gpu_p
 +
#SBATCH --ntasks=1
 +
#SBATCH --cpus-per-task=6
 +
#SBATCH --gres=gpu:P100:1
 +
#SBATCH --mem=60gb
 +
#SBATCH --constraint=Intel 
 +
#SBATCH --time=120:00:00
 +
#SBATCH --output=%x.%j.out   
 +
#SBATCH --error=%x.%j.err         
 +
 
 +
cd $SLURM_SUBMIT_DIR
 +
 
 +
ml purge
 +
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
 +
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0
 +
 
 +
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.1
 +
 
 +
singularity exec -B /apps/db/AlphaFold -B /apps/eb/CUDAcore/11.2.1 \
 +
--nv /apps/singularity-images/alphafold_2.3.1_cuda112.sif python /app/alphafold/run_alphafold.py \
 +
--use_gpu_relax \
 +
--data_dir=$ALPHAFOLD_DATA_DIR \
 +
--uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta \
 +
--mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa \
 +
--bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt \
 +
--uniref30_database_path=$ALPHAFOLD_DATA_DIR/uniref30/UniRef30_2021_03 \
 +
--pdb_seqres_database_path=$ALPHAFOLD_DATA_DIR/pdb_seqres/pdb_seqres.txt \
 +
--template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files \
 +
--obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
 +
--uniprot_database_path=$ALPHAFOLD_DATA_DIR/uniprot/uniprot.fasta \
 +
--model_preset=multimer \
 +
--max_template_date=2022-10-01 \
 +
--db_preset=full_dbs \
 +
--output_dir=./output \
 +
--fasta_paths=./input.fa
 
</pre>
 
</pre>
 +
 +
'''Notes about the singularity container for version 2.3.1:
 +
*Use the '''-B /apps/db/AlphaFold''' option to allow singularity to access the location where the database files are installed.
 +
 +
*Use the '''-B /apps/eb/CUDAcore/11.2.1''' option to allow singularity to access the CUDA libraries.
 +
 +
*Use the '''--nv''' option to allow singularity to run on a GPU. Note that the job will also need to request a GPU device using the #SBATCH --gres parameter.
 +
 +
*The only parameter for the run_alphafold.py script that you need to change in these sample job submission scripts is the path to your fasta file: --fasta_paths=
 +
 +
*You can also change these: --max_template_date and --output_dir
 +
 +
*The lines that have $ALPHAFOLD_DATA_DIR can be used exactly as they are.
 +
 +
*The job will run initially on CPU only, at a later stage it runs on a single GPU (so it suffices to request one GPU device for the job.
 +
 +
*This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).
 +
  
  
−
Sample job submission script (sub.sh) to run AlphaFold 2.0.0 using run_alphafold.sh in a batch job (with GPU):
+
Sample job submission script to run the singularity container for v. 2.2.4 for '''Monomer''' on a GPU:
 +
 
 +
<pre class="gscript">
 +
#!/bin/bash
 +
#SBATCH --job-name=alphafoldjobname     
 +
#SBATCH --partition=gpu_p       
 +
#SBATCH --ntasks=1                 
 +
#SBATCH --cpus-per-task=4
 +
#SBATCH --gres=gpu:P100:1
 +
#SBATCH --mem=50gb       
 +
#SBATCH --constraint=Intel           
 +
#SBATCH --time=120:00:00         
 +
#SBATCH --output=%x.%j.out   
 +
#SBATCH --error=%x.%j.err         
 +
 
 +
cd $SLURM_SUBMIT_DIR
 +
 
 +
ml purge
 +
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
 +
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0
 +
 
 +
ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.2.4
 +
 
 +
singularity exec -B /apps/db/AlphaFold --nv /apps/singularity-images/alphafold_2.2.4.sif python /app/alphafold/run_alphafold.py  \
 +
--use_gpu_relax \
 +
--data_dir=$ALPHAFOLD_DATA_DIR  \
 +
--uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta  \
 +
--mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa  \
 +
--bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt  \
 +
--uniclust30_database_path=$ALPHAFOLD_DATA_DIR/uniclust30/uniclust30/UniRef30_2021_03 \
 +
--pdb70_database_path=$ALPHAFOLD_DATA_DIR/pdb70/pdb70  \
 +
--template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files  \
 +
--obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
 +
--model_preset=monomer \
 +
--max_template_date=2022-1-1 \
 +
--db_preset=full_dbs \
 +
--output_dir=./output \
 +
--fasta_paths=./input.fasta
 +
</pre>
 +
 
 +
Sample job submission script to run the singularity container for v. 2.2.4 for '''Multimer''' on a GPU:
  
 
<pre class="gscript">
 
<pre class="gscript">
Line 83: Line 244:
 
#SBATCH --ntasks=1                 
 
#SBATCH --ntasks=1                 
 
#SBATCH --cpus-per-task=4
 
#SBATCH --cpus-per-task=4
−
#SBATCH --gres=gpu:K40:1
+
#SBATCH --gres=gpu:P100:1
−
#SBATCH --mem=40gb                   
+
#SBATCH --mem=50gb       
 +
#SBATCH --constraint=Intel         
 
#SBATCH --time=120:00:00           
 
#SBATCH --time=120:00:00           
 
#SBATCH --output=%x.%j.out     
 
#SBATCH --output=%x.%j.out     
 
#SBATCH --error=%x.%j.err           
 
#SBATCH --error=%x.%j.err           
−
#SBATCH --mail-user=username@uga.edu 
 
−
#SBATCH --mail-type=ALL 
 
  
 
cd $SLURM_SUBMIT_DIR
 
cd $SLURM_SUBMIT_DIR
 +
ml purge
 +
 +
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
 +
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0
  
−
ml AlphaFold/2.0.0_conda
+
ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.2.4
  
−
bash $EBROOTALPHAFOLD/alphafold/run_alphafold.sh -d /apps/db/AlphaFold [options]
+
singularity exec -B /apps/db/AlphaFold --nv /apps/singularity-images/alphafold_2.2.4.sif python /app/alphafold/run_alphafold.py \
 +
--use_gpu_relax \
 +
--data_dir=$ALPHAFOLD_DATA_DIR \
 +
--uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta \
 +
--mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa \
 +
--bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt \
 +
--uniclust30_database_path=$ALPHAFOLD_DATA_DIR/uniclust30/uniclust30/UniRef30_2021_03 \
 +
--pdb_seqres_database_path=$ALPHAFOLD_DATA_DIR/pdb_seqres/pdb_seqres.txt \
 +
--template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files \
 +
--obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
 +
--uniprot_database_path=$ALPHAFOLD_DATA_DIR/uniprot/uniprot.fasta \
 +
--model_preset=multimer \
 +
--max_template_date=2022-10-01 \
 +
--db_preset=full_dbs \
 +
--output_dir=./output \
 +
--fasta_paths=./input.fasta
  
 
</pre>
 
</pre>
  
−
where $EBROOTALPHAFOLD is the environmental variable that stores the AlphaFold installation path on the cluster; [options] need to be replaced by the options (command and arguments) you want to use. Other parameters of the job, such as the maximum wall clock time, maximum memory, the number of cores per node, and the job name need to be modified appropriately as well. You can also request a P100 device, using <code>#SBATCH --gres=gpu:P100:1</code> if you submit the job to the gpu_p partition.
+
'''Notes about the singularity container for version 2.2.4:
 +
*Use the '''-B /apps/db/AlphaFold''' to allow singularity to access the location where the database files are installed.
 +
 
 +
*Use the '''--nv''' option to allow singularity to run on a GPU. Note that the job will also need to request a GPU device using the #SBATCH --gres parameter.
 +
 
 +
*The only parameter for the run_alphafold.py script that you need to change in these sample job submission scripts is the path to your fasta file: --fasta_paths=
 +
 
 +
*You can also change these: --max_template_date and --output_dir
 +
 
 +
*The lines that have $ALPHAFOLD_DATA_DIR can be used exactly as they are.
 +
 
 +
*The job will run initially on CPU only, at a later stage it runs on a single GPU (so it suffices to request one GPU device for the job.
 +
 
 +
*This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).  
 +
 
 +
 
  
−
Sample job submission script (sub.sh) to run AlphaFold 2.0.1 in a batch job (with GPU):
+
Sample job submission script (sub.sh) to run AlphaFold 2.3.1 in a batch job (with GPU):
  
 
<pre class="gscript">
 
<pre class="gscript">
Line 108: Line 302:
 
#SBATCH --partition=gpu_p         
 
#SBATCH --partition=gpu_p         
 
#SBATCH --ntasks=1                 
 
#SBATCH --ntasks=1                 
−
#SBATCH --cpus-per-task=4
+
#SBATCH --cpus-per-task=10
−
#SBATCH --gres=gpu:P100:1
+
#SBATCH --gres=gpu:A100:1
 
#SBATCH --mem=40gb                     
 
#SBATCH --mem=40gb                     
 
#SBATCH --time=120:00:00           
 
#SBATCH --time=120:00:00           
 
#SBATCH --output=%x.%j.out     
 
#SBATCH --output=%x.%j.out     
 
#SBATCH --error=%x.%j.err           
 
#SBATCH --error=%x.%j.err           
−
#SBATCH --mail-user=username@uga.edu 
 
−
#SBATCH --mail-type=ALL 
 
  
 
cd $SLURM_SUBMIT_DIR
 
cd $SLURM_SUBMIT_DIR
  
−
ml AlphaFold/2.0.1-fosscuda-2020b
+
ml AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0
−
 
 
−
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold
 
  
 
alphafold [options]
 
alphafold [options]
Line 129: Line 319:
 
where [options] need to be replaced by the options (command and arguments) you want to use. Other parameters of the job, such as the maximum wall clock time, maximum memory, the number of cores per node, and the job name need to be modified appropriately as well.  
 
where [options] need to be replaced by the options (command and arguments) you want to use. Other parameters of the job, such as the maximum wall clock time, maximum memory, the number of cores per node, and the job name need to be modified appropriately as well.  
  
 +
An example of the options to use for the alphafold script:
 +
<pre class="gscript">
 +
alphafold --data_dir /apps/db/AlphaFold/2.3.1 --output_dir ./output --model_names model_1 --fasta_paths ./query.fasta --max_template_date 2021-11-17
 +
</pre>
  
 
Example of job submission
 
Example of job submission
Line 139: Line 333:
 
Details and references are at https://github.com/deepmind/alphafold.
 
Details and references are at https://github.com/deepmind/alphafold.
  
 +
'''Version 2.3.4:''' Short help options
 
<pre  class="gcommand">
 
<pre  class="gcommand">
−
ml AlphaFold/2.0.0_conda
+
ml AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold
  
−
bash $EBROOTALPHAFOLD/alphafold/run_alphafold.sh -h
+
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.4
  
−
Usage: /apps/gb/AlphaFold/2.0.0_conda/alphafold/run_alphafold.sh <OPTIONS>
+
alphafold --helpshort
−
Required Parameters:
 
−
-d <data_dir>    Path to directory of supporting data
 
−
-o <output_dir>  Path to a directory that will store the results.
 
−
-m <model_names>  Names of models to use (a comma separated list)
 
−
-f <fasta_path>  Path to a FASTA file containing one sequence
 
−
-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:
 
−
-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')
 
−
-g <use_gpu>      Enable NVIDIA runtime to run with GPUs (default: 'True')
 
−
-a <gpu_devices>  Comma separated list of devices to pass to 'CUDA_VISIBLE_DEVICES' (default: 'all')
 
−
-p <preset>      Choose preset model configuration - no ensembling (full_dbs) or 8 model ensemblings (casp14) (default: 'full_dbs')
 
  
 +
Full AlphaFold protein structure prediction script.
 +
flags:
 +
 +
/apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold/bin/alphafold:
 +
  --[no]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')
 +
  --bfd_database_path: Path to the BFD database for use by HHblits.
 +
    (default: '/apps/db/AlphaFold/2.3.4/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt')
 +
  --data_dir: Path to directory of supporting data.
 +
    (default: '/apps/db/AlphaFold/2.3.4')
 +
  --db_preset: <full_dbs|reduced_dbs>: Choose preset MSA database configuration - smaller genetic database config (reduced_dbs) or full genetic database config
 +
    (full_dbs)
 +
    (default: 'full_dbs')
 +
  --fasta_paths: Paths to FASTA files, each containing a prediction target that will be folded one after another. If a FASTA file contains multiple sequences,
 +
    then it will be folded as a multimer. Paths should be separated by commas. All FASTA paths must have a unique basename as the basename is used to name the
 +
    output directories for each prediction.
 +
    (a comma separated list)
 +
  --hhblits_binary_path: Path to the HHblits executable.
 +
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhblits')
 +
  --hhsearch_binary_path: Path to the HHsearch executable.
 +
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhsearch')
 +
  --hmmbuild_binary_path: Path to the hmmbuild executable.
 +
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmbuild')
 +
  --hmmsearch_binary_path: Path to the hmmsearch executable.
 +
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmsearch')
 +
  --jackhmmer_binary_path: Path to the JackHMMER executable.
 +
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/jackhmmer')
 +
  --kalign_binary_path: Path to the Kalign executable.
 +
    (default: '/apps/eb/Kalign/3.3.5-GCCcore-11.3.0/bin/kalign')
 +
  --max_template_date: Maximum template release date to consider. Important if folding historical test sets.
 +
  --mgnify_database_path: Path to the MGnify database for use by JackHMMER.
 +
    (default: '/apps/db/AlphaFold/2.3.4/mgnify/mgy_clusters_2022_05.fa')
 +
  --model_preset: <monomer|monomer_casp14|monomer_ptm|multimer>: Choose preset model configuration - the monomer model, the monomer model with extra ensembling,
 +
    monomer model with pTM head, or multimer model
 +
    (default: 'monomer')
 +
  --models_to_relax: <all|best|none>: The models to run the final relaxation step on. If `all`, all models are relaxed, which may be time consuming. If `best`,
 +
    only the most confident model is relaxed. If `none`, relaxation is not run. Turning off relaxation might result in predictions with distracting
 +
    stereochemical violations but might help in case you are having issues with the relaxation stage.
 +
    (default: 'best')
 +
  --num_multimer_predictions_per_model: How many predictions (each with a different random seed) will be generated per model. E.g. if this is 2 and there are 5
 +
    models then there will be 10 predictions per input. Note: this FLAG only applies if model_preset=multimer
 +
    (default: '5')
 +
    (an integer)
 +
  --obsolete_pdbs_path: Path to file containing a mapping from obsolete PDB IDs to the PDB IDs of their replacements.
 +
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/obsolete.dat')
 +
  --output_dir: Path to a directory that will store the results.
 +
  --pdb70_database_path: Path to the PDB70 database for use by HHsearch.
 +
    (default: '/apps/db/AlphaFold/2.3.4/pdb70/pdb70')
 +
  --pdb_seqres_database_path: Path to the PDB seqres database for use by hmmsearch.
 +
  --random_seed: The random seed for the data pipeline. By default, this is randomly generated. Note that even if this is set, Alphafold may still not be
 +
    deterministic, because processes like GPU inference are nondeterministic.
 +
    (an integer)
 +
  --small_bfd_database_path: Path to the small version of BFD used with the "reduced_dbs" preset.
 +
  --template_mmcif_dir: Path to a directory with template mmCIF structures, each named <pdb_id>.cif
 +
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/mmcif_files')
 +
  --uniprot_database_path: Path to the Uniprot database for use by JackHMMer.
 +
  --uniref30_database_path: Path to the UniRef30 database for use by HHblits.
 +
    (default: '/apps/db/AlphaFold/2.3.4/uniref30/UniRef30_2021_03')
 +
  --uniref90_database_path: Path to the Uniref90 database for use by JackHMMER.
 +
    (default: '/apps/db/AlphaFold/2.3.4/uniref90/uniref90.fasta')
 +
  --[no]use_gpu_relax: Whether to relax on GPU. Relax on GPU can be much faster than CPU, so it is recommended to enable if possible. GPUs must be available if
 +
    this setting is enabled.
 +
    (default: 'true')
 +
  --[no]use_precomputed_msas: Whether to read MSAs that have been written to disk instead of running the MSA tools. The MSA files are looked up in the output
 +
    directory, so it must stay the same between multiple runs that are to reuse the MSAs. WARNING: This will not check if the sequence, database or
 +
    configuration have changed.
 +
    (default: 'false')
 +
 +
Try --helpfull to get a list of all flags.
 +
 +
</pre>
 +
 +
'''Version 2.3.4:''' Full help options
 +
<pre  class="gcommand">
 +
ml AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold
 +
 +
export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.4
 +
 +
alphafold --helpfull
 +
 +
Full AlphaFold protein structure prediction script.
 +
flags:
 +
 +
/apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold/bin/alphafold:
 +
  --[no]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')
 +
  --bfd_database_path: Path to the BFD database for use by HHblits.
 +
    (default: '/apps/db/AlphaFold/2.3.4/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt')
 +
  --data_dir: Path to directory of supporting data.
 +
    (default: '/apps/db/AlphaFold/2.3.4')
 +
  --db_preset: <full_dbs|reduced_dbs>: Choose preset MSA database configuration - smaller genetic database config (reduced_dbs) or full genetic database config
 +
    (full_dbs)
 +
    (default: 'full_dbs')
 +
  --fasta_paths: Paths to FASTA files, each containing a prediction target that will be folded one after another. If a FASTA file contains multiple sequences,
 +
    then it will be folded as a multimer. Paths should be separated by commas. All FASTA paths must have a unique basename as the basename is used to name the
 +
    output directories for each prediction.
 +
    (a comma separated list)
 +
  --hhblits_binary_path: Path to the HHblits executable.
 +
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhblits')
 +
  --hhsearch_binary_path: Path to the HHsearch executable.
 +
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhsearch')
 +
  --hmmbuild_binary_path: Path to the hmmbuild executable.
 +
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmbuild')
 +
  --hmmsearch_binary_path: Path to the hmmsearch executable.
 +
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmsearch')
 +
  --jackhmmer_binary_path: Path to the JackHMMER executable.
 +
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/jackhmmer')
 +
  --kalign_binary_path: Path to the Kalign executable.
 +
    (default: '/apps/eb/Kalign/3.3.5-GCCcore-11.3.0/bin/kalign')
 +
  --max_template_date: Maximum template release date to consider. Important if folding historical test sets.
 +
  --mgnify_database_path: Path to the MGnify database for use by JackHMMER.
 +
    (default: '/apps/db/AlphaFold/2.3.4/mgnify/mgy_clusters_2022_05.fa')
 +
  --model_preset: <monomer|monomer_casp14|monomer_ptm|multimer>: Choose preset model configuration - the monomer model, the monomer model with extra ensembling,
 +
    monomer model with pTM head, or multimer model
 +
    (default: 'monomer')
 +
  --models_to_relax: <all|best|none>: The models to run the final relaxation step on. If `all`, all models are relaxed, which may be time consuming. If `best`,
 +
    only the most confident model is relaxed. If `none`, relaxation is not run. Turning off relaxation might result in predictions with distracting
 +
    stereochemical violations but might help in case you are having issues with the relaxation stage.
 +
    (default: 'best')
 +
  --num_multimer_predictions_per_model: How many predictions (each with a different random seed) will be generated per model. E.g. if this is 2 and there are 5
 +
    models then there will be 10 predictions per input. Note: this FLAG only applies if model_preset=multimer
 +
    (default: '5')
 +
    (an integer)
 +
  --obsolete_pdbs_path: Path to file containing a mapping from obsolete PDB IDs to the PDB IDs of their replacements.
 +
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/obsolete.dat')
 +
  --output_dir: Path to a directory that will store the results.
 +
  --pdb70_database_path: Path to the PDB70 database for use by HHsearch.
 +
    (default: '/apps/db/AlphaFold/2.3.4/pdb70/pdb70')
 +
  --pdb_seqres_database_path: Path to the PDB seqres database for use by hmmsearch.
 +
  --random_seed: The random seed for the data pipeline. By default, this is randomly generated. Note that even if this is set, Alphafold may still not be
 +
    deterministic, because processes like GPU inference are nondeterministic.
 +
    (an integer)
 +
  --small_bfd_database_path: Path to the small version of BFD used with the "reduced_dbs" preset.
 +
  --template_mmcif_dir: Path to a directory with template mmCIF structures, each named <pdb_id>.cif
 +
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/mmcif_files')
 +
  --uniprot_database_path: Path to the Uniprot database for use by JackHMMer.
 +
  --uniref30_database_path: Path to the UniRef30 database for use by HHblits.
 +
    (default: '/apps/db/AlphaFold/2.3.4/uniref30/UniRef30_2021_03')
 +
  --uniref90_database_path: Path to the Uniref90 database for use by JackHMMER.
 +
    (default: '/apps/db/AlphaFold/2.3.4/uniref90/uniref90.fasta')
 +
  --[no]use_gpu_relax: Whether to relax on GPU. Relax on GPU can be much faster than CPU, so it is recommended to enable if possible. GPUs must be available if
 +
    this setting is enabled.
 +
    (default: 'true')
 +
  --[no]use_precomputed_msas: Whether to read MSAs that have been written to disk instead of running the MSA tools. The MSA files are looked up in the output
 +
    directory, so it must stay the same between multiple runs that are to reuse the MSAs. WARNING: This will not check if the sequence, database or
 +
    configuration have changed.
 +
    (default: 'false')
 +
 +
absl.app:
 +
  -?,--[no]help: show this help
 +
    (default: 'false')
 +
  --[no]helpfull: show full help
 +
    (default: 'false')
 +
  --[no]helpshort: show this help
 +
    (default: 'false')
 +
  --[no]helpxml: like --helpfull, but generates XML output
 +
    (default: 'false')
 +
  --[no]only_check_args: Set to true to validate args and exit.
 +
    (default: 'false')
 +
  --[no]pdb: Alias for --pdb_post_mortem.
 +
    (default: 'false')
 +
  --[no]pdb_post_mortem: Set to true to handle uncaught exceptions with PDB post mortem.
 +
    (default: 'false')
 +
  --profile_file: Dump profile information to a file (for python -m pstats). Implies --run_with_profiling.
 +
  --[no]run_with_pdb: Set to true for PDB debug mode
 +
    (default: 'false')
 +
  --[no]run_with_profiling: Set to true for profiling the script. Execution will be slower, and the output format might change over time.
 +
    (default: 'false')
 +
  --[no]use_cprofile_for_profiling: Use cProfile instead of the profile module for profiling. This has no effect unless --run_with_profiling is set.
 +
    (default: 'true')
 +
 +
absl.logging:
 +
  --[no]alsologtostderr: also log to stderr?
 +
    (default: 'false')
 +
  --log_dir: directory to write logfiles into
 +
    (default: '')
 +
  --logger_levels: Specify log level of loggers. The format is a CSV list of `name:level`. Where `name` is the logger name used with `logging.getLogger()`, and
 +
    `level` is a level name  (INFO, DEBUG, etc). e.g. `myapp.foo:INFO,other.logger:DEBUG`
 +
    (default: '')
 +
  --[no]logtostderr: Should only log to stderr?
 +
    (default: 'false')
 +
  --[no]showprefixforinfo: If False, do not prepend prefix to info messages when it's logged to stderr, --verbosity is set to INFO level, and python logging is
 +
    used.
 +
    (default: 'true')
 +
  --stderrthreshold: log messages at this level, or more severe, to stderr in addition to the logfile.  Possible values are 'debug', 'info', 'warning', 'error',
 +
    and 'fatal'.  Obsoletes --alsologtostderr. Using --alsologtostderr cancels the effect of this flag. Please also note that this flag is subject to
 +
    --verbosity and requires logfile not be stderr.
 +
    (default: 'fatal')
 +
  -v,--verbosity: Logging verbosity level. Messages logged at this level or lower will be included. Set to 1 for debug logging. If the flag was not set or
 +
    supplied, the value will be changed from the default of -1 (warning) to 0 (info) after flags are parsed.
 +
    (default: '-1')
 +
    (an integer)
 +
 +
absl.testing.absltest:
 +
  --test_random_seed: Random seed for testing. Some test frameworks may change the default value of this flag between runs, so it is not appropriate for seeding
 +
    probabilistic tests.
 +
    (default: '301')
 +
    (an integer)
 +
  --test_randomize_ordering_seed: If positive, use this as a seed to randomize the execution order for test cases. If "random", pick a random seed to use. If 0
 +
    or not set, do not randomize test case execution order. This flag also overrides the TEST_RANDOMIZE_ORDERING_SEED environment variable.
 +
    (default: '')
 +
  --test_srcdir: Root of directory tree where source files live
 +
    (default: '')
 +
  --test_tmpdir: Directory for temporary testing files
 +
    (default: '/tmp/absl_testing')
 +
  --xml_output_file: File to store XML test results
 +
    (default: '')
 +
 +
tensorflow.python.ops.parallel_for.pfor:
 +
  --[no]op_conversion_fallback_to_while_loop: DEPRECATED: Flag is ignored.
 +
    (default: 'true')
 +
 +
tensorflow.python.tpu.client.client:
 +
  --[no]hbm_oom_exit: Exit the script when the TPU HBM is OOM.
 +
    (default: 'true')
 +
  --[no]runtime_oom_exit: Exit the script when the TPU runtime is OOM.
 +
    (default: 'true')
 +
 +
tensorflow.python.tpu.tensor_tracer_flags:
 +
  --delta_threshold: Log if history based diff crosses this threshold.
 +
    (default: '0.5')
 +
    (a number)
 +
  --[no]tt_check_filter: Terminate early to check op name filtering.
 +
    (default: 'false')
 +
  --[no]tt_single_core_summaries: Report single core metric and avoid aggregation.
 +
    (default: 'false')
 +
 +
absl.flags:
 +
  --flagfile: Insert flag definitions from the given file into the command line.
 +
    (default: '')
 +
  --undefok: comma-separated list of flag names that it is okay to specify on the command line even if the program does not define a flag with that name.
 +
    IMPORTANT: flags in this list that have arguments MUST use the --flag=value format.
 +
    (default: '')
  
 
</pre>
 
</pre>
Line 165: Line 583:
 
=== Installation ===
 
=== Installation ===
  
−
*Version 2.0.0: Installed using a conda environment following the steps in the dockerfile available at https://github.com/deepmind/alphafold. The run_alphafold.sh bash script was obtained from https://github.com/kalininalab/alphafold_non_docker and some documentation related to this script is available at that URL.
+
*Version 2.2.4: Installed as a singularity container pulled from https://hub.docker.com/r/catgumag/alphafold
 +
 
 +
*Version 2.3.1: Installed as a singularity container pulled from https://hub.docker.com/r/catgumag/alphafold
  
−
*Version 2.0.1: Installed using EasyBuild.
+
*Version 2.3.1: Installed using EasyBuild.
  
 
*The database files are installed in /apps/db/AlphaFold/
 
*The database files are installed in /apps/db/AlphaFold/

Latest revision as of 14:53, 25 September 2023


Category

Bioinformatics

Program On

Sapelo2

Version

2.2.4, 2.3.1, 2.3.4

Author / Distributor

Please see https://github.com/deepmind/alphafold

Description

From https://github.com/deepmind/alphafold: "This package provides an implementation of the inference pipeline of AlphaFold v2.0. This is a completely new model that was entered in CASP14 and published in Nature. "

Running Program

Also refer to Running Jobs on Sapelo2

For more information on Environment Modules on Sapelo2 please see the Lmod page.


  • Version 2.2.4

This version is installed as a singularity container pulled from https://hub.docker.com/r/catgumag/alphafold: 

/apps/singularity-images/alphafold_2.2.4.sif

You can view the documentation for this version of AlphaFold with the following command, on an interactive node: 

singularity exec /apps/singularity-images/alphafold_2.2.4.sif python /app/alphafold/run_alphafold.py --helpfull

This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).

The database files are installed in /apps/db/AlphaFold/2.2.4. To use these in the singularity container, please add the option -B /apps/db/AlphaFold to the singularity exec command, as shown in the sample job submission scripts below. The --nv option needs to be added to enable AlphaFold to run on the GPU. This singularity container also requires the option --use_gpu_relax to be added.


  • Version 2.3.1 (on Intel nodes only)

This version is installed as a singularity container pulled from https://hub.docker.com/r/catgumag/alphafold: 

/apps/singularity-images/alphafold_2.3.1_cuda112.sif

You can view the documentation for this version of AlphaFold with the following command, on an interactive node: 

singularity exec /apps/singularity-images/alphafold_2.3.1_cuda112.sif python /app/alphafold/run_alphafold.py --helpfull

This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).

The database files are installed in /apps/db/AlphaFold/2.3.1. To use these in the singularity container, please add the option -B /apps/db/AlphaFold to the singularity exec command, as shown in the sample job submission scripts below. The --nv option needs to be added to enable AlphaFold to run on the GPU. Please also add the option -B /apps/eb/CUDAcore/11.2.1 , so the singularity container can link to the CUDA libraries. This singularity container also requires the option --use_gpu_relax to be added.


  • Version 2.3.1

Installed with EasyBuild in /apps/eb/AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0/

To use this version of AlphaFold, please first load the module with 

ml AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0

Once you load the module, an environmental variable called EBROOTALPHAFOLD is exported. It stores the AlphaFold installation path on the cluster, i.e., /apps/eb/AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0. The python script run_alphafold.py is installed in EBROOTALPHAFOLD/bin and a symbolic link called alphafold points to it and can be used to run the program. The 2.2TB of database files are in /apps/db/AlphaFold/2.3.1. You can export the environment variable ALPHAFOLD_DATA_DIR to set the location of the database files. For bash, use 

export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.1

When you load the AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0 module, this environment variable will be automatically set.

Note: If you run this program on the gpu_p partition, please request a P100 or an A100 GPU device. This version requires a GPU device and it should work on the P100, V100, and A100 devices.


  • Version 2.3.4

Installed with EasyBuild in /apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold/

To use this version of AlphaFold, please first load the module with 

ml AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold

Once you load the module, an environmental variable called EBROOTALPHAFOLD is exported. It stores the AlphaFold installation path on the cluster, i.e., /apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold. The python script run_alphafold.py is installed in EBROOTALPHAFOLD/bin and a symbolic link called alphafold points to it and can be used to run the program. The 2.2TB of database files are in /apps/db/AlphaFold/2.3.4. You can export the environment variable ALPHAFOLD_DATA_DIR to set the location of the database files. For bash, use 

export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.4

When you load the AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold module, this environment variable will be automatically set.

Note: If you run this program on the gpu_p partition, please request a P100 or an A100 GPU device. This version requires a GPU device.

This version does not appear to be working now.


Sample Job Submission scripts

Sample job submission script to run the singularity container for v. 2.3.1 for Monomer on a GPU: 

#!/bin/bash
#SBATCH --job-name=alphafoldjobname       
#SBATCH --partition=gpu_p         
#SBATCH --ntasks=1                  	
#SBATCH --cpus-per-task=4
#SBATCH --gres=gpu:P100:1
#SBATCH --mem=50gb        
#SBATCH --constraint=Intel            
#SBATCH --time=120:00:00           
#SBATCH --output=%x.%j.out     
#SBATCH --error=%x.%j.err          

cd $SLURM_SUBMIT_DIR

ml purge
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0

ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.1

singularity exec -B /apps/db/AlphaFold -B /apps/eb/CUDAcore/11.2.1 \
 --nv /apps/singularity-images/alphafold_2.3.1_cuda112.sif python /app/alphafold/run_alphafold.py  \
 --use_gpu_relax \
 --data_dir=$ALPHAFOLD_DATA_DIR  \
 --uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta  \
 --mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa  \
 --bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt  \
 --uniref30_database_path=$ALPHAFOLD_DATA_DIR/uniref30/UniRef30_2021_03 \
 --pdb70_database_path=$ALPHAFOLD_DATA_DIR/pdb70/pdb70  \
 --template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files  \
 --obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
 --model_preset=monomer \
 --max_template_date=2022-10-01 \
 --db_preset=full_dbs \
 --output_dir=./output \
 --fasta_paths=./IL2Y.fasta


Sample job submission script to run the singularity container for v. 2.3.1 for Multimer on a GPU: 

#!/bin/bash
#SBATCH --job-name=alphafold
#SBATCH --partition=gpu_p
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=6
#SBATCH --gres=gpu:P100:1
#SBATCH --mem=60gb
#SBATCH --constraint=Intel  
#SBATCH --time=120:00:00
#SBATCH --output=%x.%j.out     
#SBATCH --error=%x.%j.err          

cd $SLURM_SUBMIT_DIR

ml purge
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0

export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.1

singularity exec -B /apps/db/AlphaFold -B /apps/eb/CUDAcore/11.2.1 \
--nv /apps/singularity-images/alphafold_2.3.1_cuda112.sif python /app/alphafold/run_alphafold.py \
--use_gpu_relax \
--data_dir=$ALPHAFOLD_DATA_DIR \
--uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta \
--mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa \
--bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt \
--uniref30_database_path=$ALPHAFOLD_DATA_DIR/uniref30/UniRef30_2021_03 \
--pdb_seqres_database_path=$ALPHAFOLD_DATA_DIR/pdb_seqres/pdb_seqres.txt \
--template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files \
--obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
--uniprot_database_path=$ALPHAFOLD_DATA_DIR/uniprot/uniprot.fasta \
--model_preset=multimer \
--max_template_date=2022-10-01 \
--db_preset=full_dbs \
--output_dir=./output \
--fasta_paths=./input.fa

Notes about the singularity container for version 2.3.1:

  • Use the -B /apps/db/AlphaFold option to allow singularity to access the location where the database files are installed.
  • Use the -B /apps/eb/CUDAcore/11.2.1 option to allow singularity to access the CUDA libraries.
  • Use the --nv option to allow singularity to run on a GPU. Note that the job will also need to request a GPU device using the #SBATCH --gres parameter.
  • The only parameter for the run_alphafold.py script that you need to change in these sample job submission scripts is the path to your fasta file: --fasta_paths=
  • You can also change these: --max_template_date and --output_dir
  • The lines that have $ALPHAFOLD_DATA_DIR can be used exactly as they are.
  • The job will run initially on CPU only, at a later stage it runs on a single GPU (so it suffices to request one GPU device for the job.
  • This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).


Sample job submission script to run the singularity container for v. 2.2.4 for Monomer on a GPU: 

#!/bin/bash
#SBATCH --job-name=alphafoldjobname       
#SBATCH --partition=gpu_p         
#SBATCH --ntasks=1                  	
#SBATCH --cpus-per-task=4
#SBATCH --gres=gpu:P100:1
#SBATCH --mem=50gb        
#SBATCH --constraint=Intel            
#SBATCH --time=120:00:00           
#SBATCH --output=%x.%j.out     
#SBATCH --error=%x.%j.err          

cd $SLURM_SUBMIT_DIR

ml purge
export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0

ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.2.4

singularity exec -B /apps/db/AlphaFold --nv /apps/singularity-images/alphafold_2.2.4.sif python /app/alphafold/run_alphafold.py  \
 --use_gpu_relax \
 --data_dir=$ALPHAFOLD_DATA_DIR  \
 --uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta  \
 --mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa  \
 --bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt  \
 --uniclust30_database_path=$ALPHAFOLD_DATA_DIR/uniclust30/uniclust30/UniRef30_2021_03 \
 --pdb70_database_path=$ALPHAFOLD_DATA_DIR/pdb70/pdb70  \
 --template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files  \
 --obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
 --model_preset=monomer \
 --max_template_date=2022-1-1 \
 --db_preset=full_dbs \
 --output_dir=./output \
 --fasta_paths=./input.fasta

Sample job submission script to run the singularity container for v. 2.2.4 for Multimer on a GPU: 

#!/bin/bash
#SBATCH --job-name=alphafoldjobname    
#SBATCH --partition=gpu_p         
#SBATCH --ntasks=1                  	
#SBATCH --cpus-per-task=4
#SBATCH --gres=gpu:P100:1
#SBATCH --mem=50gb         
#SBATCH --constraint=Intel           
#SBATCH --time=120:00:00           
#SBATCH --output=%x.%j.out     
#SBATCH --error=%x.%j.err          

cd $SLURM_SUBMIT_DIR
ml purge

export SINGULARITYENV_TF_FORCE_UNIFIED_MEMORY=1
export SINGULARITYENV_XLA_PYTHON_CLIENT_MEM_FRACTION=4.0

ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.2.4

singularity exec -B /apps/db/AlphaFold --nv /apps/singularity-images/alphafold_2.2.4.sif python /app/alphafold/run_alphafold.py \
--use_gpu_relax \
--data_dir=$ALPHAFOLD_DATA_DIR \
--uniref90_database_path=$ALPHAFOLD_DATA_DIR/uniref90/uniref90.fasta \
--mgnify_database_path=$ALPHAFOLD_DATA_DIR/mgnify/mgy_clusters.fa \
--bfd_database_path=$ALPHAFOLD_DATA_DIR/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt \
--uniclust30_database_path=$ALPHAFOLD_DATA_DIR/uniclust30/uniclust30/UniRef30_2021_03 \
--pdb_seqres_database_path=$ALPHAFOLD_DATA_DIR/pdb_seqres/pdb_seqres.txt \
--template_mmcif_dir=$ALPHAFOLD_DATA_DIR/pdb_mmcif/mmcif_files \
--obsolete_pdbs_path=$ALPHAFOLD_DATA_DIR/pdb_mmcif/obsolete.dat \
--uniprot_database_path=$ALPHAFOLD_DATA_DIR/uniprot/uniprot.fasta \
--model_preset=multimer \
--max_template_date=2022-10-01 \
--db_preset=full_dbs \
--output_dir=./output \
--fasta_paths=./input.fasta

Notes about the singularity container for version 2.2.4:

  • Use the -B /apps/db/AlphaFold to allow singularity to access the location where the database files are installed.
  • Use the --nv option to allow singularity to run on a GPU. Note that the job will also need to request a GPU device using the #SBATCH --gres parameter.
  • The only parameter for the run_alphafold.py script that you need to change in these sample job submission scripts is the path to your fasta file: --fasta_paths=
  • You can also change these: --max_template_date and --output_dir
  • The lines that have $ALPHAFOLD_DATA_DIR can be used exactly as they are.
  • The job will run initially on CPU only, at a later stage it runs on a single GPU (so it suffices to request one GPU device for the job.
  • This version works on nodes where the CPU processor is Intel, such as the P100 GPU nodes (note that this container does not work on the A100 node, which has an AMD processor).


Sample job submission script (sub.sh) to run AlphaFold 2.3.1 in a batch job (with GPU): 

#!/bin/bash
#SBATCH --job-name=alphafoldjobname    
#SBATCH --partition=gpu_p         
#SBATCH --ntasks=1                  	
#SBATCH --cpus-per-task=10
#SBATCH --gres=gpu:A100:1
#SBATCH --mem=40gb                    
#SBATCH --time=120:00:00           
#SBATCH --output=%x.%j.out     
#SBATCH --error=%x.%j.err          

cd $SLURM_SUBMIT_DIR

ml AlphaFold/2.3.1-foss-2022a-CUDA-11.7.0

alphafold [options]

where [options] need to be replaced by the options (command and arguments) you want to use. Other parameters of the job, such as the maximum wall clock time, maximum memory, the number of cores per node, and the job name need to be modified appropriately as well.

An example of the options to use for the alphafold script: 

alphafold --data_dir /apps/db/AlphaFold/2.3.1 --output_dir ./output --model_names model_1 --fasta_paths ./query.fasta --max_template_date 2021-11-17

Example of job submission 

sbatch sub.sh

Documentation

Details and references are at https://github.com/deepmind/alphafold.

Version 2.3.4: Short help options 

ml AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold

export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.4

alphafold --helpshort

Full AlphaFold protein structure prediction script.
flags:

/apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold/bin/alphafold:
  --[no]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')
  --bfd_database_path: Path to the BFD database for use by HHblits.
    (default: '/apps/db/AlphaFold/2.3.4/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt')
  --data_dir: Path to directory of supporting data.
    (default: '/apps/db/AlphaFold/2.3.4')
  --db_preset: <full_dbs|reduced_dbs>: Choose preset MSA database configuration - smaller genetic database config (reduced_dbs) or full genetic database config
    (full_dbs)
    (default: 'full_dbs')
  --fasta_paths: Paths to FASTA files, each containing a prediction target that will be folded one after another. If a FASTA file contains multiple sequences,
    then it will be folded as a multimer. Paths should be separated by commas. All FASTA paths must have a unique basename as the basename is used to name the
    output directories for each prediction.
    (a comma separated list)
  --hhblits_binary_path: Path to the HHblits executable.
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhblits')
  --hhsearch_binary_path: Path to the HHsearch executable.
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhsearch')
  --hmmbuild_binary_path: Path to the hmmbuild executable.
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmbuild')
  --hmmsearch_binary_path: Path to the hmmsearch executable.
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmsearch')
  --jackhmmer_binary_path: Path to the JackHMMER executable.
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/jackhmmer')
  --kalign_binary_path: Path to the Kalign executable.
    (default: '/apps/eb/Kalign/3.3.5-GCCcore-11.3.0/bin/kalign')
  --max_template_date: Maximum template release date to consider. Important if folding historical test sets.
  --mgnify_database_path: Path to the MGnify database for use by JackHMMER.
    (default: '/apps/db/AlphaFold/2.3.4/mgnify/mgy_clusters_2022_05.fa')
  --model_preset: <monomer|monomer_casp14|monomer_ptm|multimer>: Choose preset model configuration - the monomer model, the monomer model with extra ensembling,
    monomer model with pTM head, or multimer model
    (default: 'monomer')
  --models_to_relax: <all|best|none>: The models to run the final relaxation step on. If `all`, all models are relaxed, which may be time consuming. If `best`,
    only the most confident model is relaxed. If `none`, relaxation is not run. Turning off relaxation might result in predictions with distracting
    stereochemical violations but might help in case you are having issues with the relaxation stage.
    (default: 'best')
  --num_multimer_predictions_per_model: How many predictions (each with a different random seed) will be generated per model. E.g. if this is 2 and there are 5
    models then there will be 10 predictions per input. Note: this FLAG only applies if model_preset=multimer
    (default: '5')
    (an integer)
  --obsolete_pdbs_path: Path to file containing a mapping from obsolete PDB IDs to the PDB IDs of their replacements.
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/obsolete.dat')
  --output_dir: Path to a directory that will store the results.
  --pdb70_database_path: Path to the PDB70 database for use by HHsearch.
    (default: '/apps/db/AlphaFold/2.3.4/pdb70/pdb70')
  --pdb_seqres_database_path: Path to the PDB seqres database for use by hmmsearch.
  --random_seed: The random seed for the data pipeline. By default, this is randomly generated. Note that even if this is set, Alphafold may still not be
    deterministic, because processes like GPU inference are nondeterministic.
    (an integer)
  --small_bfd_database_path: Path to the small version of BFD used with the "reduced_dbs" preset.
  --template_mmcif_dir: Path to a directory with template mmCIF structures, each named <pdb_id>.cif
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/mmcif_files')
  --uniprot_database_path: Path to the Uniprot database for use by JackHMMer.
  --uniref30_database_path: Path to the UniRef30 database for use by HHblits.
    (default: '/apps/db/AlphaFold/2.3.4/uniref30/UniRef30_2021_03')
  --uniref90_database_path: Path to the Uniref90 database for use by JackHMMER.
    (default: '/apps/db/AlphaFold/2.3.4/uniref90/uniref90.fasta')
  --[no]use_gpu_relax: Whether to relax on GPU. Relax on GPU can be much faster than CPU, so it is recommended to enable if possible. GPUs must be available if
    this setting is enabled.
    (default: 'true')
  --[no]use_precomputed_msas: Whether to read MSAs that have been written to disk instead of running the MSA tools. The MSA files are looked up in the output
    directory, so it must stay the same between multiple runs that are to reuse the MSAs. WARNING: This will not check if the sequence, database or
    configuration have changed.
    (default: 'false')

Try --helpfull to get a list of all flags.

Version 2.3.4: Full help options 

ml AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold

export ALPHAFOLD_DATA_DIR=/apps/db/AlphaFold/2.3.4

alphafold --helpfull

Full AlphaFold protein structure prediction script.
flags:

/apps/eb/AlphaFold/2.3.4-foss-2022a-CUDA-11.7.0-ColabFold/bin/alphafold:
  --[no]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')
  --bfd_database_path: Path to the BFD database for use by HHblits.
    (default: '/apps/db/AlphaFold/2.3.4/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt')
  --data_dir: Path to directory of supporting data.
    (default: '/apps/db/AlphaFold/2.3.4')
  --db_preset: <full_dbs|reduced_dbs>: Choose preset MSA database configuration - smaller genetic database config (reduced_dbs) or full genetic database config
    (full_dbs)
    (default: 'full_dbs')
  --fasta_paths: Paths to FASTA files, each containing a prediction target that will be folded one after another. If a FASTA file contains multiple sequences,
    then it will be folded as a multimer. Paths should be separated by commas. All FASTA paths must have a unique basename as the basename is used to name the
    output directories for each prediction.
    (a comma separated list)
  --hhblits_binary_path: Path to the HHblits executable.
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhblits')
  --hhsearch_binary_path: Path to the HHsearch executable.
    (default: '/apps/eb/HH-suite/3.3.0-gompi-2022a/bin/hhsearch')
  --hmmbuild_binary_path: Path to the hmmbuild executable.
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmbuild')
  --hmmsearch_binary_path: Path to the hmmsearch executable.
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/hmmsearch')
  --jackhmmer_binary_path: Path to the JackHMMER executable.
    (default: '/apps/eb/HMMER/3.3.2-gompi-2022a/bin/jackhmmer')
  --kalign_binary_path: Path to the Kalign executable.
    (default: '/apps/eb/Kalign/3.3.5-GCCcore-11.3.0/bin/kalign')
  --max_template_date: Maximum template release date to consider. Important if folding historical test sets.
  --mgnify_database_path: Path to the MGnify database for use by JackHMMER.
    (default: '/apps/db/AlphaFold/2.3.4/mgnify/mgy_clusters_2022_05.fa')
  --model_preset: <monomer|monomer_casp14|monomer_ptm|multimer>: Choose preset model configuration - the monomer model, the monomer model with extra ensembling,
    monomer model with pTM head, or multimer model
    (default: 'monomer')
  --models_to_relax: <all|best|none>: The models to run the final relaxation step on. If `all`, all models are relaxed, which may be time consuming. If `best`,
    only the most confident model is relaxed. If `none`, relaxation is not run. Turning off relaxation might result in predictions with distracting
    stereochemical violations but might help in case you are having issues with the relaxation stage.
    (default: 'best')
  --num_multimer_predictions_per_model: How many predictions (each with a different random seed) will be generated per model. E.g. if this is 2 and there are 5
    models then there will be 10 predictions per input. Note: this FLAG only applies if model_preset=multimer
    (default: '5')
    (an integer)
  --obsolete_pdbs_path: Path to file containing a mapping from obsolete PDB IDs to the PDB IDs of their replacements.
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/obsolete.dat')
  --output_dir: Path to a directory that will store the results.
  --pdb70_database_path: Path to the PDB70 database for use by HHsearch.
    (default: '/apps/db/AlphaFold/2.3.4/pdb70/pdb70')
  --pdb_seqres_database_path: Path to the PDB seqres database for use by hmmsearch.
  --random_seed: The random seed for the data pipeline. By default, this is randomly generated. Note that even if this is set, Alphafold may still not be
    deterministic, because processes like GPU inference are nondeterministic.
    (an integer)
  --small_bfd_database_path: Path to the small version of BFD used with the "reduced_dbs" preset.
  --template_mmcif_dir: Path to a directory with template mmCIF structures, each named <pdb_id>.cif
    (default: '/apps/db/AlphaFold/2.3.4/pdb_mmcif/mmcif_files')
  --uniprot_database_path: Path to the Uniprot database for use by JackHMMer.
  --uniref30_database_path: Path to the UniRef30 database for use by HHblits.
    (default: '/apps/db/AlphaFold/2.3.4/uniref30/UniRef30_2021_03')
  --uniref90_database_path: Path to the Uniref90 database for use by JackHMMER.
    (default: '/apps/db/AlphaFold/2.3.4/uniref90/uniref90.fasta')
  --[no]use_gpu_relax: Whether to relax on GPU. Relax on GPU can be much faster than CPU, so it is recommended to enable if possible. GPUs must be available if
    this setting is enabled.
    (default: 'true')
  --[no]use_precomputed_msas: Whether to read MSAs that have been written to disk instead of running the MSA tools. The MSA files are looked up in the output
    directory, so it must stay the same between multiple runs that are to reuse the MSAs. WARNING: This will not check if the sequence, database or
    configuration have changed.
    (default: 'false')

absl.app:
  -?,--[no]help: show this help
    (default: 'false')
  --[no]helpfull: show full help
    (default: 'false')
  --[no]helpshort: show this help
    (default: 'false')
  --[no]helpxml: like --helpfull, but generates XML output
    (default: 'false')
  --[no]only_check_args: Set to true to validate args and exit.
    (default: 'false')
  --[no]pdb: Alias for --pdb_post_mortem.
    (default: 'false')
  --[no]pdb_post_mortem: Set to true to handle uncaught exceptions with PDB post mortem.
    (default: 'false')
  --profile_file: Dump profile information to a file (for python -m pstats). Implies --run_with_profiling.
  --[no]run_with_pdb: Set to true for PDB debug mode
    (default: 'false')
  --[no]run_with_profiling: Set to true for profiling the script. Execution will be slower, and the output format might change over time.
    (default: 'false')
  --[no]use_cprofile_for_profiling: Use cProfile instead of the profile module for profiling. This has no effect unless --run_with_profiling is set.
    (default: 'true')

absl.logging:
  --[no]alsologtostderr: also log to stderr?
    (default: 'false')
  --log_dir: directory to write logfiles into
    (default: '')
  --logger_levels: Specify log level of loggers. The format is a CSV list of `name:level`. Where `name` is the logger name used with `logging.getLogger()`, and
    `level` is a level name  (INFO, DEBUG, etc). e.g. `myapp.foo:INFO,other.logger:DEBUG`
    (default: '')
  --[no]logtostderr: Should only log to stderr?
    (default: 'false')
  --[no]showprefixforinfo: If False, do not prepend prefix to info messages when it's logged to stderr, --verbosity is set to INFO level, and python logging is
    used.
    (default: 'true')
  --stderrthreshold: log messages at this level, or more severe, to stderr in addition to the logfile.  Possible values are 'debug', 'info', 'warning', 'error',
    and 'fatal'.  Obsoletes --alsologtostderr. Using --alsologtostderr cancels the effect of this flag. Please also note that this flag is subject to
    --verbosity and requires logfile not be stderr.
    (default: 'fatal')
  -v,--verbosity: Logging verbosity level. Messages logged at this level or lower will be included. Set to 1 for debug logging. If the flag was not set or
    supplied, the value will be changed from the default of -1 (warning) to 0 (info) after flags are parsed.
    (default: '-1')
    (an integer)

absl.testing.absltest:
  --test_random_seed: Random seed for testing. Some test frameworks may change the default value of this flag between runs, so it is not appropriate for seeding
    probabilistic tests.
    (default: '301')
    (an integer)
  --test_randomize_ordering_seed: If positive, use this as a seed to randomize the execution order for test cases. If "random", pick a random seed to use. If 0
    or not set, do not randomize test case execution order. This flag also overrides the TEST_RANDOMIZE_ORDERING_SEED environment variable.
    (default: '')
  --test_srcdir: Root of directory tree where source files live
    (default: '')
  --test_tmpdir: Directory for temporary testing files
    (default: '/tmp/absl_testing')
  --xml_output_file: File to store XML test results
    (default: '')

tensorflow.python.ops.parallel_for.pfor:
  --[no]op_conversion_fallback_to_while_loop: DEPRECATED: Flag is ignored.
    (default: 'true')

tensorflow.python.tpu.client.client:
  --[no]hbm_oom_exit: Exit the script when the TPU HBM is OOM.
    (default: 'true')
  --[no]runtime_oom_exit: Exit the script when the TPU runtime is OOM.
    (default: 'true')

tensorflow.python.tpu.tensor_tracer_flags:
  --delta_threshold: Log if history based diff crosses this threshold.
    (default: '0.5')
    (a number)
  --[no]tt_check_filter: Terminate early to check op name filtering.
    (default: 'false')
  --[no]tt_single_core_summaries: Report single core metric and avoid aggregation.
    (default: 'false')

absl.flags:
  --flagfile: Insert flag definitions from the given file into the command line.
    (default: '')
  --undefok: comma-separated list of flag names that it is okay to specify on the command line even if the program does not define a flag with that name.
    IMPORTANT: flags in this list that have arguments MUST use the --flag=value format.
    (default: '')

Back to Top

Installation

  • Version 2.3.1: Installed using EasyBuild.
  • The database files are installed in /apps/db/AlphaFold/

System

64-bit Linux

Retrieved from "https://wiki.gacrc.uga.edu/index.php?title=AlphaFold-Sapelo2&oldid=21534"