examples/alphafold3.md
... ...
@@ -1,52 +1,33 @@
1 1
## AlphaFold3 Example
2 2
3
+In order to run AlphaFold3 with SBGrid you must:
4
+1. Obtain the AF3 models
5
+2. Download the AF3 databases
6
+
7
+### Obtain the AF3 models.
8
+
9
+SBGrid cannot distribute these. You must apply through this form:
10
+
11
+https://forms.gle/svvpY4u2jsHEwWYS6
12
+
13
+The link for the form is also given in the [git repository](https://github.com/google-deepmind/alphafold3)
14
+
15
+You will receive a link to download the models as a zstd compressed file. Uncompress this file to produce af3.bin. Place this file in a directory of your choice. You must supply this directory as an argument to the run_alphafold.py script.
16
+
3 17
### Download databases
4 18
5
-The databases for AlphaFOld3 (AF3) must be downloaded. We provide the script we used below. Adjust the DATABASE_DIR as needed.
19
+The databases for AlphaFold3 (AF3) must be downloaded.
6 20
21
+AlphaFold provides a script for downloading these in the github repo:
7 22
8
-```
9
-#!/usr/bin/env bash
10
-
11
-## Alphafold 3.0.0 database download
12
-## Pulled from provided fetch_databases.py
13
-##
14
-## JV Nov 21, 2024
15
-## help@sbgrid.org
16
-
17
-# get zstd in path
18
-export PATH=$PATH:$HOME/.local/bin
19
-
20
-files=(
21
- 'bfd-first_non_consensus_sequences.fasta.zst'
22
- 'mgy_clusters_2022_05.fa.zst'
23
- 'nt_rna_2023_02_23_clust_seq_id_90_cov_80_rep_seq.fasta.zst'
24
- 'pdb_2022_09_28_mmcif_files.tar.zst'
25
- 'pdb_seqres_2022_09_28.fasta.zst'
26
- 'rfam_14_9_clust_seq_id_90_cov_80_rep_seq.fasta.zst'
27
- 'rnacentral_active_seq_id_90_cov_80_linclust.fasta.zst'
28
- 'uniprot_all_2021_04.fa.zst'
29
- 'uniref90_2022_05.fa.zst'
30
-)
31
-
32
-DATABASE_DIR="/programs/local/alphafold-3.0.0/databases"
33
-GOOGLE_URL="https://storage.googleapis.com/alphafold-databases/v3.0"
34
-
35
-for filename in "${files[@]}"; do
36
- echo "Fetching: $filename"
37
- echo "curl --progress-bar --continue-at - --output ${DATABASE_DIR}/${filename} ${GOOGLE_URL}/${filename} "
38
-done
39
-
40
-for filename in "${files[@]}"; do
41
- echo "Uncompressing: $filename"
42
- echo "zstd --decompress --force ${DATABASE_DIR}/${filename}"
43
-done
23
+https://github.com/google-deepmind/alphafold3/blob/main/fetch_databases.sh
44 24
45
-```
25
+The version of AF3 curated in SBGrid uses the tar file version of the mmcif files in pdb_2022_09_28_mmcif_files.tar.
46 26
27
+Since then AF3 has changed to use the uncompressed files in the mmcif tar file. The download script will untar the file.
47 28
48 29
49
-The resulting layout:
30
+The resulting layout should look like this:
50 31
51 32
```
52 33
/programs//local/alphafold-3.0.0/databases/