Fix FABind, DynamicBind, and RFAA Conda environments

Author: amorehead

README.md

@@ -78,7 +78,7 @@ conda activate forks/DynamicBind/DynamicBind/  # NOTE: one still needs to use `c
 mamba env create -f environments/neuralplexer_environment.yaml --prefix forks/NeuralPLexer/NeuralPLexer/
 conda activate forks/NeuralPLexer/NeuralPLexer/  # NOTE: one still needs to use `conda` to (de)activate environments
 cd forks/NeuralPLexer/ && pip3 install -e . && cd ../../
-# - RoseTTAFold-All-Atom environment (~14 GB)
+# - RoseTTAFold-All-Atom environment (~14 GB) - NOTE: after running these commands, follow the installation instructions in `forks/RoseTTAFold-All-Atom/README.md` starting at Step 4 (with `forks/RoseTTAFold-All-Atom/` as the current working directory)
 mamba env create -f environments/rfaa_environment.yaml --prefix forks/RoseTTAFold-All-Atom/RFAA/
 conda activate forks/RoseTTAFold-All-Atom/RFAA/  # NOTE: one still needs to use `conda` to (de)activate environments
 cd forks/RoseTTAFold-All-Atom/rf2aa/SE3Transformer/ && pip3 install --no-cache-dir -r requirements.txt && python3 setup.py install && cd ../../../../

environments/dynamicbind_environment.yaml

@@ -239,7 +239,7 @@ dependencies:
       - platformdirs==2.5.2
       - prompt-toolkit==3.0.36
       - psutil==5.9.8
-      - git+https://github.com/pyg-team/pyg-lib.git
+      - https://data.pyg.org/whl/torch-2.1.0%2Bcu118/pyg_lib-0.4.0%2Bpt21cu118-cp39-cp39-linux_x86_64.whl
       - pygments==2.15.1
       - pyopenssl==23.0.0
       - python-dotenv==1.0.1

environments/fabind_environment.yaml

@@ -131,7 +131,7 @@ dependencies:
       - pyarrow==15.0.0
       - pyasn1==0.5.1
       - pyasn1-modules==0.3.0
-      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/pyg_lib-0.2.0%2Bpt112cu113-cp39-cp39-linux_x86_64.whl
+      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/pyg_lib-0.2.0%2Bpt112cu113-cp38-cp38-linux_x86_64.whl
       - pyparsing==3.1.1
       - python-dateutil==2.8.2
       - python-dotenv==1.0.1
@@ -149,12 +149,11 @@ dependencies:
       - tensorboard==2.14.0
       - tensorboard-data-server==0.7.2
       - threadpoolctl==3.2.0
-      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_cluster-1.6.0%2Bpt112cu113-cp39-cp39-linux_x86_64.whl
+      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_cluster-1.6.0%2Bpt112cu113-cp38-cp38-linux_x86_64.whl
       - torch-geometric==2.4.0
-      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_scatter-2.1.0%2Bpt112cu113-cp39-cp39-linux_x86_64.whl
-      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_sparse-0.6.15%2Bpt112cu113-cp39-cp39-linux_x86_64.whl
-      - torch-spline-conv==1.2.1+pt112cu113
-      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_spline_conv-1.2.1%2Bpt112cu113-cp39-cp39-linux_x86_64.whl
+      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_scatter-2.1.0%2Bpt112cu113-cp38-cp38-linux_x86_64.whl
+      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_sparse-0.6.15%2Bpt112cu113-cp38-cp38-linux_x86_64.whl
+      - https://data.pyg.org/whl/torch-1.12.0%2Bcu113/torch_spline_conv-1.2.1%2Bpt112cu113-cp38-cp38-linux_x86_64.whl
       - torchdrug==0.1.2
       - torchmetrics==0.10.2
       - tqdm==4.66.1

environments/rfaa_environment.yaml

@@ -1,5 +1,6 @@
 name: RFAA
 channels:
+  - predector
   - pyg
   - bioconda
   - pytorch
@@ -368,6 +369,7 @@ dependencies:
   - scikit-learn=1.4.1.post1=py310h1fdf081_0
   - send2trash=1.8.3=pyh0d859eb_0
   - setuptools=69.1.1=pyhd8ed1ab_0
+  - signalp6=6.0g=1
   - sip=6.7.12=py310hc6cd4ac_0
   - six=1.16.0=pyh6c4a22f_0
   - smirnoff99frosst=1.1.0=pyh44b312d_0
@@ -481,7 +483,6 @@ dependencies:
       - scipy==1.13.0
       - sentry-sdk==1.41.0
       - shortuuid==1.0.12
-      - signalp6==6.0+h
       - smmap==5.0.1
       - subprocess32==3.5.4
       - timeout-decorator==0.5.0

forks/RoseTTAFold-All-Atom/README.md

@@ -3,7 +3,7 @@ Code for RoseTTAFold All-Atom
 <p align="right">
   <img style="float: right" src="./img/RFAA.png" alt="alt text" width="600px" align="right"/>
 </p>
-RoseTTAFold All-Atom is a biomolecular structure prediction neural network that can predict a broad range of biomolecular assemblies including proteins, nucleic acids, small molecules, covalent modifications and metals as outlined in the RFAA paper. 
+RoseTTAFold All-Atom is a biomolecular structure prediction neural network that can predict a broad range of biomolecular assemblies including proteins, nucleic acids, small molecules, covalent modifications and metals as outlined in the <a href='https://www.science.org/doi/10.1126/science.adl2528'>RFAA paper</a>. 
 
 RFAA is not accurate for all cases, but produces useful error estimates to allow users to identify accurate predictions. Below are the instructions for setting up and using the model. 
 
@@ -54,16 +54,11 @@ mv $CONDA_PREFIX/lib/python3.10/site-packages/signalp/model_weights/distilled_mo
 ```
 bash install_dependencies.sh
 ```
-6. Add BLAST patch
-```
-wget https://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/blast-2.2.26-x64-linux.tar.gz
-tar -zxvf blast-2.2.26-x64-linux.tar.gz
-```
-6. Download the model weights.
+6. Download the model weights (if not already downloaded)
 ```
 wget http://files.ipd.uw.edu/pub/RF-All-Atom/weights/RFAA_paper_weights.pt
 ```
-7. Download sequence databases for MSA and template generation.
+7. Download sequence databases for MSA and template generation
 ```
 # uniref30 [46G]
 wget http://wwwuser.gwdg.de/~compbiol/uniclust/2020_06/UniRef30_2020_06_hhsuite.tar.gz
@@ -79,7 +74,17 @@ tar xfz bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt.tar.gz -C ./bfd
 wget https://files.ipd.uw.edu/pub/RoseTTAFold/pdb100_2021Mar03.tar.gz
 tar xfz pdb100_2021Mar03.tar.gz
 ```
+**NOTE:** Make sure to update `DB_UR30` and `DB_BFD` (on Lines 19 and 20 of `make_msa.sh`) as well as `database_params.hhdb` (on Line 6 of `rf2aa/config/inference/base.yaml`) to list the absolute (base) paths to these respective local databases. For example, one may set these values to `DB_UR30="/bmlfast/rfaa_databases/uniref30/UniRef30_2020_06"`, `DB_BFD="/bmlfast/rfaa_databases/bfd/bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt"`, and `hhdb: "/bmlfast/rfaa_databases/pdb100_2021Mar03/pdb100_2021Mar03"`.
 
+8. Download `BLAST`
+```
+wget https://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/blast-2.2.26-x64-linux.tar.gz
+mkdir -p blast-2.2.26
+tar -xf blast-2.2.26-x64-linux.tar.gz -C blast-2.2.26
+cp -r blast-2.2.26/blast-2.2.26/ blast-2.2.26_bk
+rm -r blast-2.2.26
+mv blast-2.2.26_bk/ blast-2.2.26
+```
 <a id="inference-config"></a>
 ### Inference Configs Using Hydra
 
@@ -150,27 +155,28 @@ python -m rf2aa.run_inference --config-name nucleic_acid
 <a id="p-sm-complex"></a>
 ### Predicting Protein Small Molecule Complexes
 To predict protein small molecule complexes, the syntax to input the protein remains the same. Adding in the small molecule works similarly to other inputs. 
-Here is an example (from `rf2aa/config/inference/protein_complex_sm.yaml`):
+Here is an example (from `rf2aa/config/inference/protein_sm.yaml`):
 ```
 defaults:
   - base
-
-job_name: 7qxr
+job_name: "3fap"
 
 protein_inputs:
-  A: 
-    fasta_file: examples/protein/7qxr.fasta
+  A:
+    fasta_file: examples/protein/3fap_A.fasta
+  B: 
+    fasta_file: examples/protein/3fap_B.fasta
 
 sm_inputs:
-  B:
-    input: examples/small_molecule/NSW_ideal.sdf
+  C:
+    input: examples/small_molecule/ARD_ideal.sdf
     input_type: "sdf"
 ```
 Small molecule inputs are provided as sdf files or smiles strings and users are **required** to provide both an input and an input_type field for every small molecule that they want to provide. Metal ions can also be provided as sdf files or smiles strings. 
 
 To predict the example:
 ```
-python -m rf2aa.run_inference --config-name protein_complex_sm
+python -m rf2aa.run_inference --config-name protein_sm
 ```
 <a id="higher-order"></a>
 ### Predicting Higher Order Complexes

forks/RoseTTAFold-All-Atom/rf2aa/config/inference/protein_sm.yaml

@@ -1,24 +1,13 @@
 defaults:
   - base
-job_name: "T1188"
+
+job_name: 7qxr
 
 protein_inputs:
-  A:
-    fasta_file: examples/protein/T1188_A.fasta
+  A: 
+    fasta_file: examples/protein/7qxr.fasta
 
 sm_inputs:
   B:
-    input: CN1CNC2C1C(O)N(CCCN1C(O)C3C(NCN3C)N(C)C1O)C(O)N2C
-    input_type: "smiles"
-  C:
-    input: Cn1cnc2c1c(=O)n(CCCn1c(=O)c3c(ncn3C)n(C)c1=O)c(=O)n2C
-    input_type: "smiles"
-  D:
-    input: [Cd+2]
-    input_type: "smiles"
-  E:
-    input: [Cd]
-    input_type: "smiles"
-  F:
-    input: [Co]
-    input_type: "smiles"
+    input: examples/small_molecule/NSW_ideal.sdf
+    input_type: "sdf"