Find open-source science resources
A directory of tools, AI models, datasets, and research resources for biotech, bioinformatics, and other scientific fields. Aggregated from curated GitHub awesome-lists, HuggingFace, bio.tools, Bioconductor, and more.
Filters
Health
Domain(1)
Language
License
Source
Type(1)
36 of 9,874 resources
Computational framework for protein design, targeting a diversity of protein design and engineering tasks, and it can also be used for protein structure modeling and scoring.
SPPIDER-seq is a partner-aware framework that predicts protein-protein interaction (PPI) sites, designed to handle both structured interfaces and disordered/motif-driven interactions by combining ESM-2 embeddings with a cross-attention architecture.
A graph attention network for structure based predictions of protein-protein interactions. A PyTorch implementation of GCN with mutual attention for protein-protein interaction prediction. Struct2Graph uses two graph convolutional networks (GCNs) with weight sharing, and a mutual attention network to extract relevant geometric features related toquery protein pairs. The details of the GCN architecture is described in our paper (currently under review).
A containerized web server for the identification and topology prediction of ABC proteins. ABC-Finder i.e., A Docker-based package for the identification of ABC proteins in all organisms, and downstream analysis and visualization of the topology of ABC proteins using an interactive web browser.
Prediction of Protein-Protein Interactions Based on L1-Regularized Logistic Regression and Gradient Tree Boosting. The code is the implementation of GTB-PPI method “Prediction of Protein-Protein Interactions Based on L1-Regularized Logistic Regression and Gradient Tree Boosting”. The dataset file contains the S. cerevisiae, H. pylori ,the indepedent data set and PPIs network (one-core network and crossover network).
HyPPI classifies a protein-protein complex based on its interaction type into permanent, transient, or crystal artifact. Permanent protein-protein complexes are only stable in their complexed state. Their subunits would denature upon dissociation of the protein-protein complex. Transient protein-protein complexes are stable in the complexed as well as in the monomeric form, depending on the necessary function of the complex. Crystal artifacts have no biological function and are artificially formed during the crystallization process. The discrimination is performed using two characteristics of the protein-protein complex, the hydrophobicity of the interface (ΔGhydrophobic) and the quotient of interface area ratios (IF-quotient). The IF-quotient considers whether the protein-protein interface is symmetric.
PoseEdit automatically generates 2D diagrams of protein-ligand complexes, focusing on the interactions between protein and ligand. Interactions between molecules are estimated by an underlying interaction model that relies on atom types and simple geometric criteria. The structure mining tool GeoMine also uses this model to describe binding sites. In addition, users can manipulate the diagrams by translating, rotating, mirroring parts of the structure, adding additional interactions, or removing them. Furthermore, users can add individual labels or adjust available labels. Users can download the final 2D diagrams for a binding site of interest in JSON or SVG format.
METALizer predicts the coordination geometry of metal ions in metalloproteins. Users can compare potential coordination geometries to those found in the examined structure. The predicted coordination geometries and the observed metal interaction distances can be interactively compared to statistics calculated based on the PDB.
PoseView automatically generates 2D diagrams of protein-ligand complexes, focusing on the interactions between protein and ligand. Interactions between molecules are estimated by an underlying interaction mode that relies on atom types and simple geometric criteria. It adheres to the conventions of chemical structure diagram generation. The quality of the resulting diagrams is comparable to manually drawn examples from books and scientific publications.
GeoMine enables the automated mining of protein-ligand binding sites. Based on individually designed queries, users can search for spatial interaction patterns in huge collections of protein-ligand complexes and binding pockets. The regularly updated GeoMine database relies on the free database systems SQLite and PostgreSQL. It supports radius-based pockets (based on ligands and predicted pockets (based on DoGSite3) for query generation. The query management is based on XML (for the REST service) or JSON in the GUI mode. Its output consists of the query-based superpositions of the matched binding sites and statistics on matching points, distances, and angles.
Protoss is a fully automated hydrogen atom placement tool for protein-ligand complexes. It adds missing hydrogen atoms to protein structures and detects reasonable protonation states, tautomeric states, and hydrogen coordinates of both protein and ligand molecules by optimizing the hydrogen bond network.
DPEB is a multimodal database that integrates four types of protein embeddings (AlphaFold2, BioEmbeddings, ESM-2, ProtVec) for advanced systems biology predictions.
PRIorI is a web-based platform designed to explore protein-protein interaction (PPI) networks by modeling them as bipartite graphs. It allows users to find similar interaction patterns within structural complexes from PDB structures.
PPLM is a protein pair language model that learns inter-protein context directly from paired sequences. It enables computational modeling of Protein-Protein Interactions (PPIs), predicting binary interaction, binding affinity, and interface contact.
ProModb (Protein interaction-based functional Module database) is a repository that documents computationally generated whole-cell protein-protein interaction network-based functional modules of an organism.
ProMoCell (Protein interaction-based functional Modules of the Cell of an organism) is a network-based zoning approach that can determine the functional modules of a cell of an organism and can potentially be utilized for parallel whole-cell simulation. ProMoCell is a single-click web service and it is very simple, user-friendly and easy to use. Presumably, no other web services like ProMoCell exists till date.
A Patient Representation with Pairwise Relative Expression of Proteins on Biological Networks. Alterations in protein and gene expression levels are often used as features to predictive models such as clinical outcome prediction.
Predicting the Effects of Single Mutations on Protein-RNA Interactions. PremPRI - Predicting the Effects of Mutations on. PremPRI predicts the effects of single muta-tions occurring in RNA binding proteins on the protein-RNA interactions by calculating the binding affinity changes quantitatively. It can be used for finding functionally importantvariants, understanding the molecular mechanisms, and designing new protein-RNA interaction inhibitors. The 3D structure of a protein-RNA complex is required forperforming the prediction.
PremPLI predicts the effects of single mutations on protein-ligand interactions by calculating the binding affinity changes quantitatively. It can be used for guiding the design of ligand-binding proteins, identifying and understanding disease driver mutations, and finding potential resistance mutations for different drugs. The 3D structure of a protein-ligand complex is required for performing the prediction.
Merging affected genes into integrated networks - multiple seeds: Seed-centric module discovery using multiple seeds.
AutoCoEv is a high-throughput in silico pipeline for predicting novel protein-protein interactions. It is a user-friendly computational pipeline for the search of coevolution between hundreds and even thousands of proteins.
A knowledge-based approach for identifying protein-nucleic acid interaction hot spots. A webserver for identifying protein-nucleic acid interfaces hotspot residues. iPNHOT was developed by using support vector machine based on three different kinds of features including atom number of the residue, depth and protrusion index, solvent accessible surface area, electrostatic potential and secondary structure. It needs one input file: pdbfile which contains the coordinates of a protein-nucleic acid complex. More details can be found in the tutorial.
Secondary Structure based End-to-End Learning model for Protein-Ligand Interaction Prediction. Abstract Computational prediction of bioactivity has become a critical aspect of modern drug discovery as it mitigates the cost, time, and resources required to find and screen new compounds. Deep Neural Networks (DNN) have recently shown excellent performance in modeling Protein-Ligand Interaction (PLI). However, DNNs are only effective when physically sound descriptions of ligands and proteins are fed into the network for further processing. Furthermore, previous research has not incorporated the secondary structure of the protein in a meaningful manner. In this work, we utilize secondary structure information of the protein which is extracted as the curvature and torsion of the backbone of protein to predict PLI. https://github.com/ekraka/SSnet
a Cytoscape application for protein-protein interaction analysis. An application for protein-protein interaction analysis. StrongestPath is a Cytoscape 3 application that enables to look for one or more cascades of interactions connecting two single or groups of proteins in a collection of protein-protein interaction (PPI) network or signaling network databases.
Regularised Automatic Prediction of Protein-Protein Interactions using Deep Learning
PIPENN stands for 'Protein Interface Prediction from sequence with an Ensemble of Neural Nets'. PIPENN is a suite of DL methods for predicting protein bindings of different interaction types (protein-protein, protein-small-molecule, protein-nucleotide(DNA/RNA)) at residue-level, using only information from a protein sequence as input.
EGRET is a deep learning based method for Protein-protein interaction site prediction, where we have introduced a novel edge aggregated graph attention network to effectively leverage the structural information.
PINOT is a Protein Interaction Network Online Tool.
An in silico approach to identification, categorization and prediction of nucleic acid binding proteins. PNIDB-the database of protein-nucleic acid interactions. PNIDB is database of protein-nucleic acid binding interactions identified by protein structure on the Protein Data Bank(PDB), as well as a collection of functional classified nucleic binding proteins based on ligand type and mmCIF keywords, each entry in PNIDB contains a comprehensive list of following annotations:. As the chief actors within the cells, protein’s interaction with nucleic acid involves many vital activities that are extremely important in the cellular process, such as transcription, translation, and DNA repair,therefore, the study on nucleic acid-protein binding activities can help to uncover the network or even the mechanism of related cellular process.
PHISDetector a tool to detect diverse in silico phage-host interaction signals for virome studies. Phage-host interactions are appealing systems to study bacterial adaptive evolution and are increasingly recognized as playing an important role in human health and diseases, which may contribute to novel therapeutic agents, such as phage therapy to combat multi-drug resistant infections. PHISDetector, a web tool to detect and systematically study diverse in silico phage-host interaction signals, including analyses for Sequence composition, CRISPR, Prophage, Genetic homology and protein-protein interactions. PHISDetector receives bacterial or virus genomic sequences in GenBank or FASTA format as input.
Mutation effect estimation on protein-protein interactions using deep contextualized representation learning. This is the repository for "Mutation effect estimation on protein-protein interactions using deep contextualized representation learning" (MuPIPR). This repository contains the source code and links to some datasets used in our paper. (to be updated)
Knowledge guided multi-level network inference. KiMONo is a network inference tool for multi-omics datasets. The multi-omics network is calculated by sparse-group-LASSO regression and can incorporate prior biological information using protein-protein Interactions.
DELPHI (DEep Learning Prediction of Highly probable protein Interaction sites) is a deep learning program for protein-protein interaction site prediction.
D-SCRIPT is a deep learning method for predicting a physical interaction between two proteins given just their sequences. It generalizes well to new species and is robust to limitations in training data size. Its design reflects the intuition that for two proteins to physically interact, a subset of amino acids from each protein should be in con-tact with the other. The intermediate stages of D-SCRIPT directly implement this intuition, with the penultimate stage in D-SCRIPT being a rough estimate of the inter-protein contact map of the protein dimer. This structurally-motivated design enhances the interpretability of the results and, since structure is more conserved evolutionarily than sequence, improves generalizability across species.
Robust Inference of Kinase Activity Using Functional Networks. RoKAI is a computation tool for inferring kinase activity in a robust manner using functional networks. RoKAI operates on a heterogeneous network having kinases and phosphosites as nodes and available functional associations as edges, including protein-protein interactions, kinase-substrate annotations, co-evolution and structure distance evidence between phosphosites. The key idea of RoKAI is to propagate the phosphosite quantifications on this heterogeneous network to capture the coordinated changes in the signaling, which are used to infer the kinase activities in a more robust manner.