Foldingdiff

Visit Tool

foldingdiff is an open-source research tool that generates novel protein backbone structures using diffusion models. It leverages trigonometry and attention mechanisms for protein structure generation.

Claim this tool

2Views

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is foldingdiff?

foldingdiff is an open-source research tool developed by Microsoft that utilizes diffusion models to generate novel protein backbone structures. It employs trigonometry and attention mechanisms, as detailed in its preprint on arXiv. The tool provides a trained model on HuggingFace spaces and SuperBio, allowing users to generate protein structures directly from a browser. It supports installation via conda and pip, and offers scripts for training custom models, downloading necessary data, and sampling protein backbones. Additionally, foldingdiff includes functionalities for evaluating designability through inverse folding with ProteinMPNN or ESM-IF1, and structural prediction using OmegaFold or AlphaFold2.

Best used for

Ideal for researchers and computational biologists who need to generate novel protein backbone structures, train custom diffusion models, and evaluate the designability of generated proteins. Especially valuable for academic research in protein science and drug discovery.

Common actions

generate protein structures

train diffusion models

evaluate protein design

visualize protein folding

workflowsautomated workflowopen-sourcedeepfakeface swappingcollaboration"AI Agents"github copilotlow-code/no-code

Capabilities

Key features

Protein backbone generation
Diffusion model training
Protein structure sampling
Inverse folding evaluation
Structural prediction
PDB file visualization

Target Audience

professorresearchercomputational biologist

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

What are the primary dependencies for installing foldingdiff?

foldingdiff is primarily written in Python and relies on PyTorch, PyTorch Lightning, and the HuggingFace transformers library. Installation is managed via a conda environment defined in the `environment.yml` file, requiring conda or mamba to set up the necessary dependencies.

Can I train my own protein generation model using foldingdiff?

Yes, foldingdiff provides scripts and configuration files to train your own diffusion model on datasets like CATH. You can customize training parameters using JSON config files and monitor the process, with model checkpoints saved in the results folder.

How does foldingdiff evaluate the designability of generated protein backbones?

foldingdiff evaluates designability by using inverse folding models like ProteinMPNN or ESM-IF1 to generate amino acid sequences for the designed backbones. These sequences are then passed through structural prediction tools like OmegaFold or AlphaFold2 to check if they fold into structures comparable to the original generated backbone.

Trending

Subcategories trending in Research & Education

Study Assistants Knowledge Management Course Creation Scientific Computing Summarization Language Learning

Trending

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce