{"id":13561105,"url":"https://github.com/HeliXonProtein/OmegaFold","last_synced_at":"2025-04-03T16:32:06.267Z","repository":{"id":49516657,"uuid":"516457982","full_name":"HeliXonProtein/OmegaFold","owner":"HeliXonProtein","description":"OmegaFold Release Code","archived":false,"fork":false,"pushed_at":"2022-12-12T14:47:33.000Z","size":734,"stargazers_count":551,"open_issues_count":44,"forks_count":78,"subscribers_count":13,"default_branch":"main","last_synced_at":"2024-11-04T12:40:23.622Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/HeliXonProtein.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2022-07-21T17:09:41.000Z","updated_at":"2024-10-26T01:28:45.000Z","dependencies_parsed_at":"2023-01-27T21:01:58.828Z","dependency_job_id":null,"html_url":"https://github.com/HeliXonProtein/OmegaFold","commit_stats":null,"previous_names":[],"tags_count":4,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HeliXonProtein%2FOmegaFold","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HeliXonProtein%2FOmegaFold/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HeliXonProtein%2FOmegaFold/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/HeliXonProtein%2FOmegaFold/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/HeliXonProtein","download_url":"https://codeload.github.com/HeliXonProtein/OmegaFold/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247037099,"owners_count":20873095,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-08-01T13:00:52.555Z","updated_at":"2025-04-03T16:32:01.257Z","avatar_url":"https://github.com/HeliXonProtein.png","language":"Python","funding_links":[],"categories":["Libraries on Molecule AI","Structure prediction","蛋白质结构","Python","Machine Learning Tasks and Models"],"sub_categories":["3D","网络服务_其他","Foundation Models"],"readme":"![header](figure.png)\n\n# OmegaFold: High-resolution de novo Structure Prediction from Primary Sequence\n\n#### This is the release code for paper [High-resolution de novo structure prediction from primary sequence](https://www.biorxiv.org/content/10.1101/2022.07.21.500999v1).\n\nWe will continue to optimize this repository for more ease of use, for\ninstance, reducing the GRAM required to inference long proteins and\nreleasing possibly stronger models.\n\n## Update Notes\n\n### Model 2 release notes Dec 9. 2022\n\nNow you can use model 2 by setting `--model 2` in the command line!\n\n### Huge GRAM reduction\n\nWe have optimized (to some extent) the GRAM usage of OmegaFold model in our\nlatest release. Now the model can inference protein sequence as long as\n_4096_ on NVIDIA A100 Graphics card with 80 GB of memory with\n`--subbatch_size` set to 448 without hitting full memory.\nThis version's model is more sensitive to `--subbatch_size`.\n\n### Setting Subbatch\n\nSubbatch makes a trade-off between time and space.\nOne can greatly reduce the space requirements by setting `--subbatch_size`\nvery low.\nThe default is the number of residues in the sequence and the lowest\npossible number is 1.\nFor now we do not have a rule of thumb for setting the `--subbatch_size`,\nbut we suggest half the value if you run into GPU memory limitations.\n\n### MacOS Users\n\nFor macOS users, we support MPS (Apple Silicon) acceleration if the user\ninstalls the latest nightly version of PyTorch.\nAlso, current code also requires macOS users need to `git clone` the\nrepository and use `python main.\npy` (see below) to run the model.\n\n## Setup\n\nTo prepare the environment to run OmegaFold,\n\n- from source\n\n```commandline\npip install git+https://github.com/HeliXonProtein/OmegaFold.git\n```\n\n- clone the repository\n\n```commandline\ngit clone https://github.com/HeliXonProtein/OmegaFold\ncd OmegaFold\npython setup.py install\n```\n\nshould get you where you want.\n\nThe `INPUT_FILE.fasta` should be a normal fasta file with possibly many\nsequences with a comment line starting with `\u003e` or `:` above the amino\nacid sequence itself.\n\nThis command will download the weight\nfrom https://helixon.s3.amazonaws.com/release1.pt\nto `~/.cache/omegafold_ckpt/model.pt`\nand load the model\n\n## Running\n\nYou could simply\n\n```commandline\nomegafold INPUT_FILE.fasta OUTPUT_DIRECTORY\n```\n\nAnd voila!\n\n### Alternatively (Or MacOS users)\n\nEven if this failed, since we use minimal 3rd party libraries, you can\nalways just install the latest\n[PyTorch](https://pytorch.org) and [biopython](https://biopython.org)\n(and that's it!) yourself.\nFor mps accelerator, macOS users may need to install the lastest nightly\nversion of PyTorch.\nIn this case, you could run\n\n```commandline\npython main.py INPUT_FILE.fasta OUTPUT_DIRECTORY\n```\n\n### Notes on resources\n\nHowever, since we have implemented sharded execution, it is possible to\n\n1. trade computation time for GRAM: by changing `--subbatch_size`. The\n   smaller\n   this value is, the longer the execution can take, and the less memory is\n   required, or,\n2. trade computation time for average prediction quality, by changing\n   `--num_cycle`\n\nFor more information, run\n\n```commandline\nomegafold --help\n```\n\nwhere we provide several options for both speed and weights utilities.\n\n## Output\n\nWe produce one pdb for each of the sequences in `INPUT_FILE.fasta` saved in\nthe `OUTPUT_DIRECTORY`. We also put our confidence value the place of\nb_factors in pdb files.\n\n## Cite\n\nIf this is helpful to you, please consider citing the paper with\n\n```tex\n@article{OmegaFold,\n\tauthor = {Wu, Ruidong and Ding, Fan and Wang, Rui and Shen, Rui and Zhang, Xiwen and Luo, Shitong and Su, Chenpeng and Wu, Zuofan and Xie, Qi and Berger, Bonnie and Ma, Jianzhu and Peng, Jian},\n\ttitle = {High-resolution de novo structure prediction from primary sequence},\n\telocation-id = {2022.07.21.500999},\n\tyear = {2022},\n\tdoi = {10.1101/2022.07.21.500999},\n\tpublisher = {Cold Spring Harbor Laboratory},\n\tURL = {https://www.biorxiv.org/content/early/2022/07/22/2022.07.21.500999},\n\teprint = {https://www.biorxiv.org/content/early/2022/07/22/2022.07.21.500999.full.pdf},\n\tjournal = {bioRxiv}\n}\n\n```\n\n## Note\n\nAlso some of the comments might be out-of-date as of now, and will be\nupdated very soon","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FHeliXonProtein%2FOmegaFold","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FHeliXonProtein%2FOmegaFold","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FHeliXonProtein%2FOmegaFold/lists"}