Merge pull request #19942 from ThomasHoffmann77/20240220124705_new_pr…

…_AlphaFold232

{bio}[foss/2023a] AlphaFold v2.3.2, dm-haiku v0.0.12, tensorstore v0.1.65 w/ CUDA v12.1.1

easybuild/easyconfigs/a/AlphaFold/AlphaFold-2.3.2-foss-2023a-CUDA-12.1.1.eb

@@ -0,0 +1,167 @@
+easyblock = 'PythonBundle'
+
+name = 'AlphaFold'
+version = '2.3.2'
+versionsuffix = '-CUDA-%(cudaver)s'
+
+homepage = 'https://deepmind.com/research/case-studies/alphafold'
+description = "AlphaFold can predict protein structures with atomic accuracy even where no similar structure is known"
+
+toolchain = {'name': 'foss', 'version': '2023a'}
+
+builddependencies = [
+    ('poetry', '1.5.1')
+]
+
+dependencies = [
+    ('Python', '3.11.3'),
+    ('CUDA', '12.1.1', '', SYSTEM),
+    ('SciPy-bundle', '2023.07'),
+    ('PyYAML', '6.0'),
+    ('TensorFlow', '2.13.0'),  # doesn't require TF-gpu
+    ('Biopython', '1.83'),
+    ('HH-suite', '3.3.0'),
+    ('HMMER', '3.4'),
+    ('Kalign', '3.4.0'),
+    ('jax', '0.4.25', versionsuffix),  # also provides absl-py  # requirement is ==0.3.25!
+    ('UCX-CUDA', '1.14.1', versionsuffix),
+    ('cuDNN', '8.9.2.26', versionsuffix, SYSTEM),
+    ('NCCL', '2.18.3', versionsuffix),
+    ('OpenMM', '8.0.0', versionsuffix),
+    ('dm-tree', '0.1.8'),
+    ('dm-haiku', '0.0.12', versionsuffix),
+]
+
+# commit to use for downloading stereo_chemical_props.txt and copy to alphafold/common,
+# see docker/Dockerfile in AlphaFold repository
+local_scp_commit = '7102c6'
+
+components = [
+    ('stereo_chemical_props.txt',  local_scp_commit, {
+        'easyblock': 'Binary',
+        'source_urls': [
+            'https://git.scicore.unibas.ch/schwede/openstructure/-/raw/%s/modules/mol/alg/src/' % local_scp_commit,
+        ],
+        'sources': [
+            {
+                'download_filename': 'stereo_chemical_props.txt',
+                'filename': 'stereo_chemical_props-%s.txt' % local_scp_commit,
+                'extract_cmd': "cp %s ./stereo_chemical_props.txt",
+            }
+        ],
+        'checksums': [
+            '24510899eeb49167cffedec8fa45363a4d08279c0c637a403b452f7d0ac09451',  # stereo_chemical_props-7102c6.txt
+        ]
+    })
+]
+
+use_pip = True
+
+exts_list = [
+    ('PDBFixer', '1.9', {
+        'source_urls': ['https://github.com/openmm/pdbfixer/archive/refs/tags/'],
+        'sources': [{'download_filename': '%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'checksums': ['88b9a77e50655f89d0eb2075093773e82c27a4cef842cb7d735c877b20cd39fb'],
+    }),
+    ('tabulate', '0.9.0', {
+        'checksums': ['0095b12bf5966de529c0feb1fa08671671b3368eec77d7ef7ab114be2c068b3c'],
+    }),
+    ('websocket-client', '1.5.1', {
+        'modulename': 'websocket',
+        'checksums': ['3f09e6d8230892547132177f575a4e3e73cfdf06526e20cc02aa1c3b47184d40'],
+    }),
+    ('docker', '7.0.0', {
+        'checksums': ['323736fb92cd9418fc5e7133bc953e11a9da04f4483f828b527db553f1e7e5a3'],
+    }),
+    ('immutabledict', '4.1.0', {
+        'checksums': ['93d100ccd2cd09a1fd3f136b9328c6e59529ba341de8bb499437f6819159fe8a'],
+    }),
+    ('contextlib2', '21.6.0', {
+        'checksums': ['ab1e2bfe1d01d968e1b7e8d9023bc51ef3509bba217bb730cee3827e1ee82869'],
+    }),
+    ('ml_collections', '0.1.1', {
+        'preinstallopts': "touch requirements.txt && touch requirements-test.txt && ",
+        'checksums': ['3fefcc72ec433aa1e5d32307a3e474bbb67f405be814ea52a2166bfc9dbe68cc'],
+    }),
+    (name, version, {
+        'patches': [
+            'AlphaFold-2.0.0_fix-packages.patch',
+            'AlphaFold-2.3.2_data-dep-paths-shebang-UniRef30.patch',
+            'AlphaFold-2.0.0_n-cpu.patch',
+            'AlphaFold-2.0.1_setup_rm_tfcpu.patch',
+            'AlphaFold-2.3.2_use_openmm_8.0.0.patch',
+            'AlphaFold-2.3.2_BioPythonPDBData.patch',
+        ],
+        'source_urls': ['https://github.com/deepmind/alphafold/archive/refs/tags/'],
+        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'checksums': [
+            {'AlphaFold-2.3.2.tar.gz': '4ea8005ba1b573fa1585e4c29b7d188c5cbfa59b4e4761c9f0c15c9db9584a8e'},
+            {'AlphaFold-2.0.0_fix-packages.patch': '826d2d1a5d6ac52c51a60ba210e1947d5631a1e2d76f8815305b5d23f74458db'},
+            {'AlphaFold-2.3.2_data-dep-paths-shebang-UniRef30.patch':
+             '58cd0ce4094afe76909649abe68034c4fbdb500967f5c818f49b530356dc012b'},
+            {'AlphaFold-2.0.0_n-cpu.patch': 'dfda4dd5f9aba19fe2b6eb9a0ec583d12dcefdfee8ab8803fc57ad48d582db04'},
+            {'AlphaFold-2.0.1_setup_rm_tfcpu.patch':
+             '1a2e4e843bd9a4d15ee39e6c37cc63ba281311cc7a0a5610f0e43b52ef93faac'},
+            {'AlphaFold-2.3.2_use_openmm_8.0.0.patch':
+             'bbef940c0c959040aaf3984ec47777a229c164517b54616a2688d58fae636d84'},
+            {'AlphaFold-2.3.2_BioPythonPDBData.patch':
+             'e4483a525ae5c4dc5a5f633bed8cf5337c329e64b603ab7b684a9d18cd26a22f'},
+        ],
+    }),
+]
+
+local_pylibdir = '%(installdir)s/lib/python%(pyshortver)s/site-packages'
+local_link_scp = 'ln -s %%(installdir)s/stereo_chemical_props.txt %s/alphafold/common' % local_pylibdir
+
+postinstallcmds = [
+    'cp %(builddir)s/AlphaFold/alphafold-%(version)s/run_alphafold*.py %(installdir)s/bin',
+    'cp -rpP %(builddir)s/AlphaFold/alphafold-%(version)s/scripts %(installdir)s',
+    'cd %(installdir)s/bin && ln -s run_alphafold.py alphafold',
+    'chmod a+x %(installdir)s/bin/run_alphafold.py',
+    local_link_scp,
+]
+
+sanity_check_paths = {
+    'files': ['bin/alphafold', 'bin/pdbfixer', 'bin/run_alphafold.py', 'stereo_chemical_props.txt'],
+    'dirs': ['lib/python%(pyshortver)s/site-packages', 'scripts'],
+}
+
+sanity_check_commands = [
+    "pdbfixer --help",
+    "python -m openmm.testInstallation",
+    "python -c 'import alphafold'",
+    "alphafold --help 2>&1 | grep 'Full AlphaFold protein structure prediction script'",
+    "python %(installdir)s/bin/run_alphafold_test.py",
+]
+
+sanity_pip_check = True
+
+# these allow to make predictions on proteins that would typically be too long to fit into GPU memory;
+# see https://github.com/deepmind/alphafold/blob/main/docker/run_docker.py
+modextravars = {
+    # these allow to make predictions on proteins that would typically be too long to fit into GPU memory;
+    # see https://github.com/deepmind/alphafold/blob/main/docker/run_docker.py
+    'TF_FORCE_UNIFIED_MEMORY': '1',
+    # jaxlib 0.4.1: https://jax.readthedocs.io/en/latest/changelog.html#jaxlib-0-4-1-dec-13-2022 
+    # "The behavior of XLA_PYTHON_CLIENT_MEM_FRACTION=.XX has been changed to allocate XX% of the total GPU memory 
+    # instead of the previous behavior of using currently available GPU memory to calculate preallocation. Please refer
+    # to GPU memory allocation for more details."
+    # https://jax.readthedocs.io/en/latest/gpu_memory_allocation.html
+    'XLA_PYTHON_CLIENT_MEM_FRACTION': '2.5',
+    # 
+    # Download with $EBROOTALPHAFOLD/scripts/download_all_data.sh /path/to/AlphaFold_DBs/$EBVERSIONALPHAFOLD
+    'ALPHAFOLD_DATA_DIR': '/path/to/AlphaFold_DBs/%(versions)s',  # please adapt
+    # Adapt in order to use a different version of UniRef30 by default,
+    # e.g.,  v2023_02 from https://wwwuser.gwdg.de/~compbiol/uniclust/2023_02/UniRef30_2023_02_hhsuite.tar.gz:
+    'ALPHAFOLD_UNIREF30_VER': '2021_03',
+    'OPENMM_RELAX': 'CUDA'  # unset or set to 'CPU' in order not to run the energy minimization on GPU; PR#189
+}
+
+postinstallmsgs = [
+    "A newer version of UniRef30 (2023_02) is available at: "
+    "https://wwwuser.gwdg.de/~compbiol/uniclust/2023_02/UniRef30_2023_02_hhsuite.tar.gz. "
+    "Untar to $ALPHAFOLD_DATA_DIR/uniref30/ and set the default version accordingly by changing "
+    "modextravars:ALPHAFOLD_UNIREF30_VER."
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/a/AlphaFold/AlphaFold-2.3.2_BioPythonPDBData.patch

@@ -0,0 +1,14 @@
+# Thomas Hoffmann, EMBL Heidelberg, [email protected], 2024/10
+# BioPython 1.83 does not provide protein_letters_3to1 in Bio.Data.SCOPdata but in Bio.Data.PDBData (and Bio.Data.IUPACData)
+diff -ru -ru alphafold-2.3.2/alphafold/data/mmcif_parsing.py alphafold-2.3.2_BioPythonSCOPData/alphafold/data/mmcif_parsing.py
+--- alphafold-2.3.2/alphafold/data/mmcif_parsing.py	2024-02-19 09:55:16.359778490 +0100
++++ alphafold-2.3.2_BioPythonSCOPData/alphafold/data/mmcif_parsing.py	2023-03-27 13:50:49.000000000 +0200
+@@ -21,7 +21,7 @@
+
+ from absl import logging
+ from Bio import PDB
+-from Bio.Data import SCOPData
++from Bio.Data import PDBData as SCOPData
+
+ # Type aliases:
+ ChainId = str

easybuild/easyconfigs/a/AlphaFold/AlphaFold-2.3.2_data-dep-paths-shebang-UniRef30.patch

@@ -0,0 +1,164 @@
+pick up on $ALPHAFOLD_DATA_DIR to specify location to downloaded data
+(see https://github.com/deepmind/alphafold/blob/main/docker/run_docker.py);
+pick up on HH-suite, HHMER, Kalign dependencies provided via EasyBuild
+author: Kenneth Hoste (HPC-UGent)
+update 2.0.1 -> 2.1.0/2.1.2/2.3.0/2.3.2: Thomas Hoffmann (EMBL);
+uniref30 version env. variable (THEMBL)
+
+diff -ru alphafold-2.3.2/run_alphafold.py alphafold-2.3.2_data-dep-paths-shebang-UniRef30/run_alphafold.py
+--- alphafold-2.3.2/run_alphafold.py	2023-03-27 13:50:49.000000000 +0200
++++ alphafold-2.3.2_data-dep-paths-shebang-UniRef30/run_alphafold.py	2024-10-11 11:34:06.330278962 +0200
+@@ -1,3 +1,4 @@
++#!/usr/bin/env python
+ # Copyright 2021 DeepMind Technologies Limited
+ #
+ # Licensed under the Apache License, Version 2.0 (the "License");
+@@ -42,6 +43,48 @@
+ import numpy as np
+
+ # Internal import (7716).
++use_reduced_dbs = any("--db_preset=reduced_dbs" in s for s in sys.argv[1:])
++use_monomer_preset = not any("--model_preset=multimer" in s for s in sys.argv[1:])
++
++data_dir = os.getenv('ALPHAFOLD_DATA_DIR')
++use_gpu_relax = os.getenv('OPENMM_RELAX')=='CUDA'
++uniref30_ver = os.getenv('ALPHAFOLD_UNIREF30_VER')
++if not uniref30_ver: uniref30_ver = '2021_03'
++
++if data_dir:
++    mgnify_database_path = os.path.join(data_dir, 'mgnify', 'mgy_clusters_2022_05.fa')
++    uniref90_database_path = os.path.join(data_dir, 'uniref90', 'uniref90.fasta')
++    template_mmcif_dir = os.path.join(data_dir, 'pdb_mmcif', 'mmcif_files')
++    obsolete_pdbs_path = os.path.join(data_dir, 'pdb_mmcif', 'obsolete.dat')
++    if use_monomer_preset:
++        pdb_seqres_database_path = None
++        uniprot_database_path = None
++        pdb70_database_path = os.path.join(data_dir, 'pdb70', 'pdb70')
++    else:
++        pdb_seqres_database_path = os.path.join(data_dir, 'pdb_seqres', 'pdb_seqres.txt')
++        uniprot_database_path = os.path.join(data_dir, 'uniprot', 'uniprot.fasta')
++        pdb70_database_path = None
++    if use_reduced_dbs:
++        small_bfd_database_path =  os.path.join(data_dir, 'small_bfd','bfd-first_non_consensus_sequences.fasta')
++        uniref30_database_path = None 
++        bfd_database_path = None 
++    else:
++        small_bfd_database_path = None 
++        bfd_database_path = os.path.join(data_dir, 'bfd', 'bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt')
++        uniref30_database_path = os.path.join(data_dir, 'uniref30', 'UniRef30_%s' % uniref30_ver)
++else:
++    sys.stderr.write("$ALPHAFOLD_DATA_DIR is not defined!")
++    uniref90_database_path = None
++    mgnify_database_path = None
++    bfd_database_path = None
++    uniref30_database_path = None
++    pdb70_database_path = None
++    template_mmcif_dir = None
++    obsolete_pdbs_path = None
++    small_bfd_database_path = None
++    uniprot_database_path = None
++    pdb_seqres_database_path = None 
++    use_gpu_relax = None
+
+ logging.set_verbosity(logging.INFO)
+
+@@ -59,7 +102,7 @@
+     'separated by commas. All FASTA paths must have a unique basename as the '
+     'basename is used to name the output directories for each prediction.')
+
+-flags.DEFINE_string('data_dir', None, 'Path to directory of supporting data.')
++flags.DEFINE_string('data_dir', data_dir, 'Path to directory of supporting data.')
+ flags.DEFINE_string('output_dir', None, 'Path to a directory that will '
+                     'store the results.')
+ flags.DEFINE_string('jackhmmer_binary_path', shutil.which('jackhmmer'),
+@@ -71,32 +114,32 @@
+ flags.DEFINE_string('hmmsearch_binary_path', shutil.which('hmmsearch'),
+                     'Path to the hmmsearch executable.')
+ flags.DEFINE_string('hmmbuild_binary_path', shutil.which('hmmbuild'),
+-                    'Path to the hmmbuild executable.')
++                     'Path to the hmmbuild executable.')
+ flags.DEFINE_string('kalign_binary_path', shutil.which('kalign'),
+-                    'Path to the Kalign executable.')
+-flags.DEFINE_string('uniref90_database_path', None, 'Path to the Uniref90 '
+-                    'database for use by JackHMMER.')
+-flags.DEFINE_string('mgnify_database_path', None, 'Path to the MGnify '
+-                    'database for use by JackHMMER.')
+-flags.DEFINE_string('bfd_database_path', None, 'Path to the BFD '
+-                    'database for use by HHblits.')
+-flags.DEFINE_string('small_bfd_database_path', None, 'Path to the small '
+-                    'version of BFD used with the "reduced_dbs" preset.')
+-flags.DEFINE_string('uniref30_database_path', None, 'Path to the UniRef30 '
+-                    'database for use by HHblits.')
+-flags.DEFINE_string('uniprot_database_path', None, 'Path to the Uniprot '
+-                    'database for use by JackHMMer.')
+-flags.DEFINE_string('pdb70_database_path', None, 'Path to the PDB70 '
+-                    'database for use by HHsearch.')
+-flags.DEFINE_string('pdb_seqres_database_path', None, 'Path to the PDB '
+-                    'seqres database for use by hmmsearch.')
+-flags.DEFINE_string('template_mmcif_dir', None, 'Path to a directory with '
+-                    'template mmCIF structures, each named <pdb_id>.cif')
++                     'Path to the Kalign executable.')
++flags.DEFINE_string('uniref90_database_path', uniref90_database_path, 'Path to the Uniref90 '
++                     'database for use by JackHMMER.')
++flags.DEFINE_string('mgnify_database_path', mgnify_database_path, 'Path to the MGnify '
++                     'database for use by JackHMMER.')
++flags.DEFINE_string('bfd_database_path', bfd_database_path, 'Path to the BFD '
++                     'database for use by HHblits.')
++flags.DEFINE_string('small_bfd_database_path', small_bfd_database_path, 'Path to the small '
++                     'version of BFD used with the "reduced_dbs" preset.')
++flags.DEFINE_string('uniref30_database_path', uniref30_database_path, 'Path to the UniRef30 '
++                     'database for use by HHblits.')
++flags.DEFINE_string('uniprot_database_path', uniprot_database_path, 'Path to the Uniprot '
++                     'database for use by JackHMMer.')
++flags.DEFINE_string('pdb70_database_path', pdb70_database_path, 'Path to the PDB70 '
++                     'database for use by HHsearch.')
++flags.DEFINE_string('pdb_seqres_database_path', pdb_seqres_database_path, 'Path to the PDB '
++                     'seqres database for use by hmmsearch.')
++flags.DEFINE_string('template_mmcif_dir', template_mmcif_dir, 'Path to a directory with '
++                     'template mmCIF structures, each named <pdb_id>.cif')
+ flags.DEFINE_string('max_template_date', None, 'Maximum template release date '
+-                    'to consider. Important if folding historical test sets.')
+-flags.DEFINE_string('obsolete_pdbs_path', None, 'Path to file containing a '
+-                    'mapping from obsolete PDB IDs to the PDB IDs of their '
+-                    'replacements.')
++                     'to consider. Important if folding historical test sets.')
++flags.DEFINE_string('obsolete_pdbs_path', obsolete_pdbs_path, 'Path to file containing a '
++                     'mapping from obsolete PDB IDs to the PDB IDs of their '
++                     'replacements.')
+ flags.DEFINE_enum('db_preset', 'full_dbs',
+                   ['full_dbs', 'reduced_dbs'],
+                   'Choose preset MSA database configuration - '
+@@ -137,7 +180,7 @@
+                         'distracting stereochemical violations but might help '
+                         'in case you are having issues with the relaxation '
+                         'stage.')
+-flags.DEFINE_boolean('use_gpu_relax', None, 'Whether to relax on GPU. '
++flags.DEFINE_boolean('use_gpu_relax', 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.')
+@@ -334,6 +377,10 @@
+                        'sure it is installed on your system.')
+
+   use_small_bfd = FLAGS.db_preset == 'reduced_dbs'
++  if use_small_bfd and data_dir:
++    bfd_database_path = None
++    uniref30_database_path = None
++
+   _check_flag('small_bfd_database_path', 'db_preset',
+               should_be_set=use_small_bfd)
+   _check_flag('bfd_database_path', 'db_preset',
+@@ -456,13 +503,7 @@
+   flags.mark_flags_as_required([
+       'fasta_paths',
+       'output_dir',
+-      'data_dir',
+-      'uniref90_database_path',
+-      'mgnify_database_path',
+-      'template_mmcif_dir',
+       'max_template_date',
+-      'obsolete_pdbs_path',
+-      'use_gpu_relax',
+   ])
+
+   app.run(main)