As also reported in #13, running the af2_interface_metrics.py step in the SH3 tutorial results in the following Jax error:

TypeError: _forward_fn() takes 1 positional argument but 2 were given

Specifically, this is the code that causes it.

 prediction_result = jax.vmap(model_runner.apply, in_axes=(None,None,0,0))(model_runner.params,
          jax.random.PRNGKey(0), processed_feature_dict, processed_initial_guess_dict)

#13 got closed by the issue author as removing the initial_guess logic allowed the code to run.

I opened this issue as it is my understanding that this initial guess logic is a big part of the method which was outlined in https://doi.org/10.1101/2022.06.15.495993 (N. Bennet et al 2022).

Do you know of a way to fix this error or is it still valuable to run it without the initial_guess? Seems to be the last step i need before being able to run through the tutorial in its entirety :)

Hello!

I am trying the inpainting test PDB. When I try the default mask or another loop, I get a garbled PDB output. I realize it's backbone-only, but it has not structure, just looks like spaghetti. Any recommendations?

python ../inpaint.py --task hal --checkpoint v02 --pdb=input/2KL8.pdb --out=output/inpaint_test --window=A,20,25 --num_designs 1 --inf_method=multi_shot --n_iters 10 --recycle_str_mode both  --clamp_1d 0.1 --dump_all

Hi, developers

i am trying to run test in inpaint and i got the following problem: Where can i find the folder and file "envsjwatson3_first_full_epoch/BFF_epoch_0_fraction_1.0.pt" Thanks a lot.

Loading checkpoint v02: /home/yanx/Dev/repos/RFDesign/envsjwatson3_first_full_epoch/BFF_epoch_0_fraction_1.0.pt... Traceback (most recent call last): File "../inpaint.py", line 837, in main() File "../inpaint.py", line 375, in main ckpt = torch.load(cp_path, map_location=DEVICE) File "/home/yanx/.conda/envs/se3test/lib/python3.8/site-packages/torch/serialization.py", line 594, in load with _open_file_like(f, 'rb') as opened_file: File "/home/yanx/.conda/envs/se3test/lib/python3.8/site-packages/torch/serialization.py", line 230, in _open_file_like return _open_file(name_or_buffer, mode) File "/home/yanx/.conda/envs/se3test/lib/python3.8/site-packages/torch/serialization.py", line 211, in init super(_open_file, self).init(open(name, mode)) FileNotFoundError: [Errno 2] No such file or directory: '/home/yanx/Dev/repos/RFDesign/envsjwatson3_first_full_epoch/BFF_epoch_0_fraction_1.0.pt'

I created a Dockerfile to run RFdesign (including af2_metrics.py) using CUDA. Getting all the versions of packages right was surprisingly tricky, so I thought I would contribute it back in case you want to add it the repository. I verified that inpainting, hallucination, and af2_metrics.py works on a V100. This does not include pyrosetta however due to the licensing constraints. Thanks for all your work on RFDesign.

SOLUTION TO THIS PROBLEM IS TO COMMENT OUT LINE 40 of does af2_metrics.py that says "from alphafold.relax import relax" (also discussed in #7) Hi RFDesign Team,

Thank you very much for sharing RFDesign and providing the tutorials!

I am trying to run the af2_metrics.py script however the SE3-nvidia.yml file does not contain a number of modules required to run it. So far I have identified the following needing to be installed: dm-tree, biopython, dm-haiku, absl-py, jaxlib, jax, ml_collections, tensorflow I got up to PDBfixer and its requirement openmm but unfortunately, openmm wouldn't install because of environment conflicts.

I notice that the same issue has been highlighted in issue #7 and so I have tried to install various different correct versions of the different libraries and I have also gone to the alpha fold git repo to identify their requirements.txt folders. Unfortunately, I am still not having any luck with getting the environment conflicts to solve :(

You mentioned in #7 that you were working on a major update to the repo that would solve this env issue but I notice that the SE3-nvidia.yml file still doesn't contain some of the above.

I was wondering if you might be able to update the yml file to build the right conda environment? Would I then be able to run the conda env -f create SE3-nvidia.yml again and it would overwrite the previous SE3-nvidia env with the new updated one?

Thank you!

Hello, I have recently read your latest paper "Accurate Mutation Effect Prediction using RoseTTAFold", a great work, and I want to try it, but I do not know how to mask out the amino acid in that mutation position in MSA or what to replace the mutant amino acid with. Could you please provide a complete use case and example data, thank you.

I have several questions about free hallucination to generate protein binders.

Here is my command line input.

python3 /opt/tools/RFDesign/hallucination/hallucinate.py 
--pdb=./binder.pdb 
--out=test/test 
--steps=m300 
--num=10 
--start_num=1 
--mask=13,A9-15,60 
--spike=0.05 
--spike_fas=./binder_80aa.fasta 
--exclude_aa=C 
--receptor=./receptor_trunc.pdb 
--rec_placement=second 
--w_rog=1 
--rog_thresh=16 
--save_pdb=True 
--track_step 1 
--use_template=A9-15

Q1 I want to know whether the --mask=10,A6-10,65 (total 80 aa) and --mask=80 are same when the argument --spike is set to 0.05.

In addition, Since the --spike=0.05 means random sequence, I think that there is no significancy which sequence is enter to the --spike_fas, but there is significancy only in the sequence length. is it right ?

Q2 In the Fig S17-C of the 2022 Science paper,

it seems that only the binding target (= receptor?) was used as template.

If it is right, how can I enter the binding target as template ? which argument is required ?

or is binding target automatically provided as template to hallucination if the binding target pdb file is provided to the --receptor argument ?

Q3 To perform free hallucination, I run the script without --use_template, but some error was occurred.

Saving /proc_1/RFDesign/binder/test/test_1: Traceback (most recent call last):
  File "/opt/tools/RFDesign/hallucination/hallucinate.py", line 739, in <module>
    main()
  File "/opt/tools/RFDesign/hallucination/hallucinate.py", line 731, in main
    optimization.save_result(out_prefix, Net, ml, trb['msa'], args, trb, 
  File "/opt/tools/RFDesign/hallucination/util/../optimization.py", line 179, in save_result
    idx_tmpl = net_kwargs['idx'].cpu().numpy()[0]
KeyError: 'idx'

When I enter the dummy value (like Z999-1000) to the --use_template, then the script worked well.

is it proper way to perform free hallucination ?

Sincerely,

jongseo

Hello dear author!

I'm very interested in your work, it's a great job!

So, I've recently been designing new sequences with RFDesign. In the process, I tried to execute RFDesign/scripts/af2_metrics.py to score the results using Alphafold2, but I ran into some problems: the process requires the AF2 dependency to be installed in the RFDesign environment, and I tried to install the AF2 dependency on top of RFDesign base to install the alphafold2 environment, but it always fails to execute af2_metrics.py successfully, the result is as follows:

I personally think it is the problem of jax andtensorflow and cuda version mismatch. What I need to ask is, what version of jax and python are you using and can you expose the files with RFDesign and alphafold2 dependencies in them? For example xx.yml or requirement.txt

This is my conda list:

Name Version Build Channel _libgcc_mutex 0.1 main conda-forge _openmp_mutex 4.5 1_gnu defaults _tflow_select 2.1.0 gpu defaults abseil-cpp 20210324.0 h9c3ff4c_0 conda-forge absl-py 1.1.0 pyhd8ed1ab_0 conda-forge aiohttp 3.7.4.post0 py38h497a2fe_0 conda-forge astor 0.8.1 pyh9f0ad1d_0 conda-forge astunparse 1.6.3 pyhd8ed1ab_0 conda-forge async-timeout 3.0.1 py_1000 conda-forge attrs 21.4.0 pyhd8ed1ab_0 conda-forge biopython 1.79 pypi_0 pypi blas 1.0 mkl conda-forge blinker 1.4 py_1 conda-forge brotlipy 0.7.0 py38h27cfd23_1003 defaults bzip2 1.0.8 h7b6447c_0 defaults c-ares 1.17.1 h7f98852_1 conda-forge ca-certificates 2022.5.18.1 ha878542_0 conda-forge cachetools 4.2.4 pyhd8ed1ab_0 conda-forge certifi 2022.5.18.1 py38h578d9bd_0 conda-forge cffi 1.14.6 py38h400218f_0 defaults chardet 4.0.0 py38h578d9bd_3 conda-forge charset-normalizer 2.0.4 pyhd3eb1b0_0 defaults chex 0.0.7 pypi_0 pypi click 8.1.3 py38h578d9bd_0 conda-forge colorama 0.4.4 pyh9f0ad1d_0 conda-forge contextlib2 21.6.0 pypi_0 pypi cryptography 36.0.0 py38h9ce1e76_0 defaults cudatoolkit 11.3.1 h2bc3f7f_2 defaults dgl-cuda11.3 0.7.2 py38_0 dglteam dm-haiku 0.0.4 pypi_0 pypi dm-tree 0.1.6 pypi_0 pypi ffmpeg 4.3 hf484d3e_0 pytorch fftw 3.3.9 h27cfd23_1 defaults freetype 2.11.0 h70c0345_0 defaults gast 0.3.3 py_0 conda-forge giflib 5.2.1 h7b6447c_0 defaults gmp 6.2.1 h2531618_2 defaults gnutls 3.6.15 he1e5248_0 defaults google-auth 1.35.0 pyh6c4a22f_0 conda-forge google-auth-oauthlib 0.4.6 pyhd8ed1ab_0 conda-forge google-pasta 0.2.0 pyh8c360ce_0 conda-forge googledrivedownloader 0.4 pyhd3deb0d_1 conda-forge grpc-cpp 1.37.1 h36de60a_0 conda-forge grpcio 1.37.1 py38hdd6454d_0 conda-forge h5py 2.10.0 nompi_py38h9915d05_106 conda-forge hdf5 1.10.6 nompi_h7c3c948_1111 conda-forge hhsuite 3.3.0 py38pl5262hc37a69a_2 bioconda hmmer 3.3.2 h1b792b2_1 bioconda icu 68.1 h58526e2_0 conda-forge idna 3.3 pyhd3eb1b0_0 defaults immutabledict 2.0.0 pypi_0 pypi importlib-metadata 4.11.4 py38h578d9bd_0 conda-forge intel-openmp 2021.4.0 h06a4308_3561 defaults jax 0.3.13 pypi_0 pypi jaxlib 0.3.10 pypi_0 pypi jinja2 3.0.3 pyhd8ed1ab_0 conda-forge joblib 1.1.0 pyhd8ed1ab_0 conda-forge jpeg 9d h7f8727e_0 defaults kalign2 2.04 h779adbc_2 bioconda keras-preprocessing 1.1.2 pyhd8ed1ab_0 conda-forge krb5 1.19.2 hcc1bbae_0 conda-forge lame 3.100 h7b6447c_0 defaults lcms2 2.12 h3be6417_0 defaults ld_impl_linux-64 2.35.1 h7274673_9 defaults libcurl 7.78.0 h2574ce0_0 conda-forge libedit 3.1.20191231 he28a2e2_2 conda-forge libev 4.33 h516909a_1 conda-forge libffi 3.3 he6710b0_2 defaults libgcc-ng 9.3.0 h5101ec6_17 defaults libgfortran-ng 7.5.0 ha8ba4b0_17 defaults libgfortran4 7.5.0 ha8ba4b0_17 defaults libgomp 9.3.0 h5101ec6_17 defaults libiconv 1.15 h63c8f33_5 defaults libidn2 2.3.2 h7f8727e_0 defaults libnghttp2 1.43.0 h812cca2_0 conda-forge libpng 1.6.37 hbc83047_0 defaults libprotobuf 3.15.8 h780b84a_0 conda-forge libssh2 1.9.0 ha56f1ee_6 conda-forge libstdcxx-ng 9.3.0 hd4cf53a_17 defaults libtasn1 4.16.0 h27cfd23_0 defaults libtiff 4.2.0 h85742a9_0 defaults libunistring 0.9.10 h27cfd23_0 defaults libuv 1.40.0 h7b6447c_0 defaults libwebp 1.2.0 h89dd481_0 defaults libwebp-base 1.2.0 h27cfd23_0 defaults lie-learn 0.0.1.post1 pypi_0 pypi lz4-c 1.9.3 h295c915_1 defaults markdown 3.3.7 pyhd8ed1ab_0 conda-forge markupsafe 2.0.1 py38h497a2fe_0 conda-forge mkl 2021.4.0 h06a4308_640 defaults mkl-service 2.4.0 py38h7f8727e_0 defaults mkl_fft 1.3.1 py38hd3c417c_0 defaults mkl_random 1.2.2 py38h51133e4_0 defaults ml-collections 0.1.0 pypi_0 pypi multidict 5.1.0 py38h497a2fe_1 conda-forge ncurses 6.3 h7f8727e_2 defaults nettle 3.7.3 hbbd107a_1 defaults networkx 2.6.3 pyhd3eb1b0_0 defaults numpy 1.21.2 py38h20f2e39_0 defaults numpy-base 1.21.2 py38h79a1101_0 defaults oauthlib 3.2.0 pyhd8ed1ab_0 conda-forge ocl-icd 2.3.0 h7f98852_0 conda-forge ocl-icd-system 1.0.0 1 conda-forge olefile 0.46 pyhd3eb1b0_0 defaults openh264 2.1.1 h4ff587b_0 defaults openmm 7.5.1 py38ha082873_1 conda-forge openssl 1.1.1o h7f8727e_0 defaults opt_einsum 3.3.0 pyhd8ed1ab_1 conda-forge packaging 21.3 pyhd3eb1b0_0 defaults pandas 1.3.1 py38h1abd341_0 conda-forge pdbfixer 1.7 pyhd3deb0d_0 conda-forge perl 5.26.2 h14c3975_0 defaults pillow 8.4.0 py38h5aabda8_0 defaults pip 21.2.4 py38h06a4308_0 defaults protobuf 3.15.8 py38h709712a_0 conda-forge pyasn1 0.4.8 py_0 conda-forge pyasn1-modules 0.2.7 py_0 conda-forge pycparser 2.21 pyhd3eb1b0_0 defaults pyg 2.0.3 py38_torch_1.10.0_cu113 pyg pyjwt 2.4.0 pyhd8ed1ab_0 conda-forge pyopenssl 21.0.0 pyhd3eb1b0_1 defaults pyparsing 3.0.4 pyhd3eb1b0_0 defaults pysocks 1.7.1 py38h06a4308_0 defaults python 3.8.12 h12debd9_0 defaults python-dateutil 2.8.2 pyhd8ed1ab_0 conda-forge python-flatbuffers 1.12 pyhd8ed1ab_1 conda-forge python-louvain 0.15 pyhd3deb0d_0 conda-forge python_abi 3.8 2_cp38 conda-forge pytorch 1.10.1 py3.8_cuda11.3_cudnn8.2.0_0 pytorch pytorch-cluster 1.5.9 py38_torch_1.10.0_cu113 pyg pytorch-mutex 1.0 cuda pytorch pytorch-scatter 2.0.9 py38_torch_1.10.0_cu113 pyg pytorch-sparse 0.6.12 py38_torch_1.10.0_cu113 pyg pytorch-spline-conv 1.2.1 py38_torch_1.10.0_cu113 pyg pytz 2021.3 pyhd8ed1ab_0 conda-forge pyu2f 0.1.5 pyhd8ed1ab_0 conda-forge pyyaml 5.4.1 py38h497a2fe_0 conda-forge re2 2021.04.01 h9c3ff4c_0 conda-forge readline 8.1 h27cfd23_0 defaults requests 2.26.0 pyhd3eb1b0_0 defaults requests-oauthlib 1.3.1 pyhd8ed1ab_0 conda-forge rsa 4.8 pyhd8ed1ab_0 conda-forge scikit-learn 1.0.1 py38h51133e4_0 defaults scipy 1.7.3 py38hc147768_0 defaults setuptools 58.0.4 py38h06a4308_0 defaults six 1.16.0 pyhd3eb1b0_0 defaults snappy 1.1.8 he1b5a44_3 conda-forge sqlite 3.37.0 hc218d9a_0 defaults svgwrite 1.4.2 pypi_0 pypi tabulate 0.8.9 pypi_0 pypi tensorboard 2.4.1 pyhd8ed1ab_1 conda-forge tensorboard-plugin-wit 1.8.1 pyhd8ed1ab_0 conda-forge tensorflow 2.4.1 py38h578d9bd_0 conda-forge tensorflow-base 2.4.1 py38h83f5f1d_0 conda-forge tensorflow-estimator 2.4.0 pyh9656e83_0 conda-forge tensorflow-gpu 2.4.1 h30adc30_0 defaults termcolor 1.1.0 py_2 conda-forge threadpoolctl 3.0.0 pyh8a188c0_0 conda-forge tk 8.6.11 h1ccaba5_0 defaults toolz 0.11.2 pypi_0 pypi torchaudio 0.10.1 py38_cu113 pytorch torchvision 0.11.2 py38_cu113 pytorch tqdm 4.62.3 pyhd8ed1ab_0 conda-forge tree 0.2.4 pypi_0 pypi typing-extensions 3.10.0.2 hd3eb1b0_0 defaults typing_extensions 3.10.0.2 pyh06a4308_0 defaults urllib3 1.26.7 pyhd3eb1b0_0 defaults werkzeug 2.1.2 pyhd8ed1ab_1 conda-forge wheel 0.37.0 pyhd3eb1b0_1 defaults wrapt 1.12.1 py38h497a2fe_3 conda-forge xz 5.2.5 h7b6447c_0 defaults yacs 0.1.6 py_0 conda-forge yaml 0.2.5 h516909a_0 conda-forge yarl 1.6.3 py38h497a2fe_2 conda-forge zipp 3.8.0 pyhd8ed1ab_0 conda-forge zlib 1.2.11 h7f8727e_4 defaults zstd 1.4.9 haebb681_0 defaults

I would really like to get your answer, thank you very much!

The af2_metrics.py requests a module called alphafold in hallucination/models/, but actually there is not such a module. I've copied to hallucination/models/, but there are still some compatible problems(maybe Deepmind updated their work?). Could you please push this module to your repository?

I think this code should maybe have a wiki attached to it to add general knowledge? I have been able to run and get at least something out. The PDBs do not inlclude rotamers/sidechains, but do include the identity of the residues. What is your general workflow for then incorporating these sequences and modeling the side chains? Do you have scripts available that you could share that does this?

Been doing a local install and going through the SH3 tutorial. I am stuck at the relaxation step after getting AF2 interface metrics:

/home/nrbennet/rosetta_builds/tutorial_build/main/source/cmake/build_release/rosetta_scripts \ 
@/home/nrbennet/protocols/flags/relax/relax.flags \ 
-parser:protocol \
/home/nrbennet/protocols/xml/relax/min/minterface.xml

Does the repo contain the flags @/home/nrbennet/protocols/flags/relax/relax.flags and the XML script /home/nrbennet/protocols/xml/relax/min/minterface.xml required to run this relaxation?

Would it be possible to provide these for the tutorial or is there another way to "complete" the tutorial?

Is there a restriction on the AA length of a protein to run RFDesign? I encounter errors using 16GB and 24GB gpus (v100 and a30)

RuntimeError: CUDA out of memory. Tried to allocate 9.36 GiB (GPU 0; 23.55 GiB total capacity; 14.56 GiB already allocated; 6.95 GiB free; 14.96 GiB reserved in total by PyTorch)

https://github.com/RosettaCommons/RFDesign/blob/bba6992283de63faba6ff727bb4bc68327a5356c/hallucination/hallucinate.py#L109

https://github.com/RosettaCommons/RFDesign/blob/bba6992283de63faba6ff727bb4bc68327a5356c/hallucination/loss.py#L680

Looking at how the code works, seems like -1 is to turn off, not 1. That's unless I'm understanding the code wrong (which is possible!).

When I calculate Rosetta interface metrics by running sbatch --mem 5g --wrap "../../scripts/get_interface_metrics.py output/run2/trf_relax/"

I get sh: /software/psipred4/runpsipred_single: No such file or directory

How can I install psipred4?

In Tutorial: Hallucinating PD-L1 binder, Pt 1, Generating Commands, force_aa is declared but never used. force_aa is declared again and used only in "Run 2: 2-chain hallucination".

This is confusing, because it's unclear whether to use force_aa or mask in order to specify which residues to keep the same. One or the other ought to make the other redundant.

When running the get_interface_metrics.py? I get the following error

sh: /software/rosetta/DAlphaBall.gcc: No such file or directory sh: /software/rosetta/DAlphaBall.gcc: No such file or directory From where should Iget the DAlphaBall?

I got the following error messages when running af2_interface_metrics.py

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "../../../../../../scripts/af2_interface_metrics.py", line 600, in <module>
    predict_structure(tag_buffer, feature_dict_dict, binderlen_dict, initial_guess_dict, sfd_out, scorefilename)
  File "../../../../../../scripts/af2_interface_metrics.py", line 429, in predict_structure
    prediction_result = jax.vmap(model_runner.apply, in_axes=(None,None,0,0))(model_runner.params, jax.random.PRNGKey(0), processed_feature_dict, processed_initial_guess_dict)
  File "/HOME/scz3220/.conda/envs/rfdesign/lib/python3.8/site-packages/haiku/_src/transform.py", line 127, in apply_fn
    out, state = f.apply(params, {}, *args, **kwargs)
  File "/HOME/scz3220/.conda/envs/rfdesign/lib/python3.8/site-packages/haiku/_src/transform.py", line 354, in apply_fn
    out = f(*args, **kwargs)
TypeError: _forward_fn() takes 1 positional argument but 2 were given

it seems like the current version of alphafold attached in this directory hallucination/models/alphafold/alphafold/ is not compatible with the scripts af2_interface_metrics.py.

May I ask how to modify alphafold scripts to pass initial guess of binders' position? Thanks for sharing this great work! and best wishes~