Tool Catalog

AbAtlas is a tool for dimensionality reduction and visualization of antibody sequences, capable of mapping antibody sequences into two-dimensional and three-dimensional graphics. This tool utilizes data from the Observed Antibody Space(OAS) database, which includes heavy and light chains from six major species (human, mouse, rat, rhesus monkey, camel, and rabbit), as well as their germline genes. By combining AntiBERTy and UMAP, AbAtlas generates high-quality sequence embeddings and effectively performs dimensionality reduction. Simply input an antibody sequence, and AbAtlas will automatically analyze the sequence and visually display its similarity to antibody chains from different species or various V gene families through graphical representations, allowing for the rapid identification of features in the input sequence.

High-concentration antibody solutions are essential for the development of subcutaneous injectable formulations, but they often exhibit high viscosity, which poses challenges to antibody drug development, production, and administration. Previous computational models have been limited to training on only a few dozen data points, which is a bottleneck for generalization.

Boltz is a fundamental model in structural biology jointly launched by institutions including MIT, focusing on the modeling of biomolecular interactions.

Protenix is an open-source model launched by ByteDance, which reproduces AlphaFold3 using PyTorch. It focuses on biomolecular structure prediction and features performance, methodology, data, and openness at the forefront of the industry:

Chai-1 is a multimodal molecular structure prediction foundation model, focusing on accurate 3D structure prediction of biomolecules. It achieves state-of-the-art performance in drug discovery and biomolecular interaction studies. Its core value lies in using deep learning techniques to decode the folded structures and interaction mechanisms of biomolecules, providing critical support for targeted drug design and protein function studies.

RFantibody utilizes RFdiffusion and RoseTTAFold2 to fine-tune the structures of natural antibodies, specifically for antibody structure design and prediction, supporting the design of single-domain antibodies (VHH). It is capable of designing antibody structures with high binding affinity based on specified antigen epitopes. The design process is as follows: * Given the antibody framework structure and the target antigen structure, binding hotspots can be specified. * Using the diffusion model technique of RFdiffusion, the antibody structure is progressively "denoised" and optimized to design CDR loops that bind to the epitopes of the target antigen. * CDR loop sequences are designed using ProteinMPNN4, achieving an amino acid recovery rate of 52.4%. * The structure of the antibody-antigen complex is predicted and screened using the fine-tuned RoseTTAFold2.

ImmuneBuilder, including ABodyBuilder2, NanoBodyBuilder2, and TCRBuilder2, is specifically designed for predicting the structure of immunoproteins (e.g., antibodies, nanobodies, and T-cell receptors), and employs AlphaFold-Multimer's structural modules with modifications specific to the immunoproteins to improve prediction accuracy. ImmuneBuilder is able to quickly generate immunoprotein structures that resemble experimental data much faster than AlphaFold2 and without the need for large sequence databases or multiple sequence comparisons. The tool's features include high accuracy, fast prediction, and open-source accessibility for structural analysis of large-scale sequence datasets, especially in the study of immunoprotein structures from next-generation sequencing data. ImmuneBuilder also provides error estimation to help filter out erroneous models, enhancing its value for applications in biotherapeutics and immunology research. Figure 1 shows the architecture of AbBuilder2, and the same architecture is used for NanoBodyBuilder2 and TCRBuilder2.

ImmuneBuilder, including ABodyBuilder2, NanoBodyBuilder2, and TCRBuilder2, is specifically designed for predicting the structure of immunoproteins (e.g., antibodies, nanobodies, and T-cell receptors), and employs AlphaFold-Multimer's structural modules with modifications specific to the immunoproteins to improve prediction accuracy.ImmuneBuilder is able to quickly generate immunoprotein structures that resemble experimental data much faster than AlphaFold2 and without the need for large sequence databases or multiple sequence comparisons. The tool's features include high accuracy, fast prediction, and open-source accessibility for structural analysis of large-scale sequence datasets, especially in the study of immunoprotein structures from next-generation sequencing data. immuneBuilder also provides error estimation to help filter out erroneous models, enhancing its value for applications in biotherapeutics and immunology research. Figure 1 shows the architecture of AbBuilder2, and the same architecture is used for NanoBodyBuilder2 and TCRBuilder2.

VDJGermline is a bioinformatics-based tool for antibody sequence VDJ analysis. Through sequence alignment, it can quickly retrieve the most homologous sequences to the target antibody sequences from a rich VDJ gene library, and deeply analyze the VDJ gene rearrangement information. It is particularly suitable for the sequence analysis of animal immune repertoires. The gene library covers multiple species, including humans, mice, rhesus monkeys, and camels, and has specifically constructed the “HUGO H3K3” fully humanized mouse gene library. It outputs analysis result tables for each sequence, visual alignment charts, and a variety of statistical analysis charts.

CDR Annotation is an antibody numbering and annotation module used to number the variable region (Fv) of antibody sequences, accurately marking the specific locations of the framework region (Framework Region, FWR) and the complementarity determining region (Complementarity Determining Region, CDR). It supports the IMGT, Kabat, Chothia, Martin, AHo and Wolfguy schemes. When multiple sequences are input at once, you can analyze sequence variations and conservation by viewing the sequence visualization and amino acid frequency plots.

The module can determine the probability that an antibody belongs to human based on its V-region sequence.

HDOCK uses a global search method based on Fast Fourier Transform (FFT) for sampling by a modified shape complementarity scoring method. During docking, one molecule (e.g. receptor) is fixed and the other molecule (e.g. ligand) is rotated uniformly in 3D Eulerian space. For each rotation of the ligand, the receptor and ligand are mapped onto a mesh and possible binding modes are exhaustively sampled in 3D translational space using the FFT method. The general case is rigid-body docking, although the flexibility problem can be handled indirectly by providing the residue information of the binding sites as constraints.

GeoDock is a novel multi-track iterative transformer network designed to address limitations in conventional protein-protein docking algorithms and existing deep learning methods. It is capable of predicting docked structures from separate docking partners, allowing for flexibility at the protein residue level to accommodate conformational changes upon binding. GeoDock attains an average inference speed of under one second on a single GPU, enabling its application in large-scale structure screening.

Nanobodies (Nbs) have recently emerged as a promising class of antibody fragments for biomedical and therapeutic applications. Despite their notable physicochemical properties, Nbs, which are derived from camelids, may require “humanization” to enhance their translational potential for clinical trials. The authors of Llamanade systematically analyzed the sequence and structural properties of Nbs based on next-generation sequencing (NGS) databases and high-resolution structures. Their analysis revealed substantial framework diversities and highlighted key differences between Nbs and human Immunoglobulin G (IgG) antibodies. They identified conserved residues that may contribute to improved solubility, structural stability, and antigen-binding, providing valuable insights into the humanization of Nbs.

AbPair is an artificial intelligence tool trained on advanced AI algorithms and a vast dataset of natural antibody pairing sequences, designed to predict the natural pairing scores of heavy and light chains. It efficiently evaluates the pairing potential between heavy and light chains, providing a scientific basis for antibody design and optimization.

A flow-based generative model (named Rotamer Density Estimator, RDE) estimates the probability distribution of conformation , uses entropy as the measure of flexibility and predict the binding ∆∆G.

Using PPB-Affinity, currently the largest protein-protein binding affinity database, as training data, the magnitude of protein complex binding affinity (ΔG) is predicted using invariant point notation based on geometric deep learning techniques through three-dimensional characterisation of protein complexes.

The thermostability of proteins is of significant importance in the biotechnology field, particularly in industries such as pharmaceuticals, food production, and biofuel generation. Thermostable proteins can accelerate chemical reactions and reduce production costs. However, traditional experimental methods for assessing protein thermostability are not only time-consuming and expensive but also difficult to scale, resulting in a limited availability of protein thermostability data.

ProteinMPNN has outstanding performance in both in silico and experimental tests. On native protein backbones, ProteinMPNN has a sequence recovery of 52.4% compared with 32.9% for Rosetta. The amino acid sequence at different positions can be coupled between single or multiple chains, enabling application to a wide range of current protein design challenges.

Protein Interaction Calculator is specifically designed to analyze the interactions between residues in protein structures. By calculating the distances between specific functional residues, it can identify and classify different types of interactions. These interactions include, but are not limited to, hydrophobic interactions, hydrogen bonds, ionic interactions, aromatic interactions, and disulfide bridges.

Calculate physicochemical properties of protein sequences.

Phylogenetic Tree takes aligned antibody sequences as input to construct a phylogenetic tree diagram, which aids in analyzing the evolutionary relationships between the sequences and reveals the origins and evolutionary processes of the antibodies. The phylogenetic inference methods include NJ (Neighbor Joining), UPGMA (Unweighted Pair Group Method with Arithmetic Mean), ME (Minimum Evolution), ML (Maximum Likelihood), and MP (Maximum Parsimony).

Multiple Sequence Alignment is used for aligning DNA and protein sequences, and visualizing the results of the sequence alignment. It aids in sequence clustering, analyzing diversity among sequences, identifying conserved regions and mutations. It includes automatic alignment tools such as ClustalW and MUSCLE, with MUSCLE incorporating clustering methods like NJ(Neighbor Joining), UPGMA(Unweighted Pair Group Method with Arithmetic Mean), and UPGMB(Unweighted Pair Group Method with Banded Mean).

Monoclonal antibody therapeutics typically originate from non-human sources (usually mice), which may trigger immune responses in humans. Antibody humanization aims to modify the variable region sequences of antibodies to obtain antibodies that do not elicit immune responses. We utilized nearly one billion antibody sequences from the OAS database to establish an antibody humanness evaluation AI model capable of distinguishing between human and non-human antibody variable region sequences. The scores output by the model are negatively correlated with the experimental immunogenicity (ADA) of existing FDA-approved antibody therapies. Following the approach of Marks and Hummer, we combined this model with a Beam Search algorithm to develop an antibody sequence humanization tool. This tool aims to maximize the level of humaness of antibodies while minimizing number of mutations and maintaining key characteristics such as affinity, thereby reducing their immunogenicity.

Prediction of absolute protein stability ΔG by protein sequence inverse folding model ESM-IF. Traditional physical methods (e.g., FoldX, Rosetta, etc.) for predicting protein stability ΔG rely on high-confidence structural pdb, and if there are too many mutations, the structural confidence decreases and the prediction results are poor. Benchmark results at ProteinGym show that the generative model ESM-IF predicts protein mutation stability ΔΔG of DMS data at best-in-class level in zero-shot. The method is an extension of mutation prediction by using the ESM-IF model to directly predict the absolute ΔG value of intact protein folding stability. It was tested with a prediction error RMSE ≈ 1.5 kcal/mol and a correlation coefficient of 0.7, representing a major breakthrough in predicting the folding stability ΔΔG of proteins.

PTM Hotspot is a tool designed to identify potential post-translational modification (PTM) risk sites within antibody sequences. By scanning sequences, it accurately locates PTM sites that may affect antibody stability, activity, or immunogenicity, providing critical data support for antibody drug development. Additionally, PTM Hotspot features powerful visualization capabilities, presenting PTM risk site information in an intuitive sequence map. It also enables users to visually assess, via scatter plots, whether sequences fall within therapeutic antibody risk regions. This helps researchers quickly interpret analysis results, optimize antibody sequence design, reduce risks in biopharmaceutical development, and accelerate the R&D process.

Antibody drug developability risk assessment and druggability analysis are critical steps in the drug discovery pipeline, aiming to identify promising clinical candidates early in the development process while mitigating potential risks. Building upon previous work (TAP tool), we developed AbTrimmer, a computational tool that evaluates antibody drug development risks based on multiple biophysical parameters, including Patches of Surface Hydrophobicity (PSH), Patches of Surface Positive Charge (PPC), Patches of Surface Negative Charge (PNC), Structural Fv charge symmetry parameter (SFvCSP), and aggregation scores. By precisely quantifying antibody features such as hydrophobicity and charge distribution, and comparing against clinically validated or marketed therapeutic antibodies, AbTrimmer enables comprehensive risk assessment of antibody molecules.

RFdiffusion is an open source method for structure generation, with or without conditional information (a motif, target etc). In a manner analogous to networks which produce images from user-specified inputs, RFdiffusion enables the design of diverse, complex, functional proteins from simple molecular specifications.

Similarity AbScan is a specialized tool for similarity searching of antibody sequence patent literature. It features a comprehensive database sourced from authoritative websites, divided into two sub-databases: Antibody and Nanobody. The Nanobody sub-database contains nearly 5,000 VHH, VNAR, and single-domain antibody sequences derived from patents and academic papers. The Antibody sub-database includes over 670,000 heavy and light chain antibody sequences, as well as 170,000 pairs of antibody sequences matched through patents and academic papers.Users can search for similar antibody sequences or structures by entering a sequence, or they can search for related antibodies by keywords. Among the paired antibody sequences, there are: * Patented Antibodies (patent in heavy definition) with 134,183 entries, accounting for 75.3%. * Crystal Structures (Xtal structure) with 12,778 entries, accounting for 7.1%. * Therapeutic Antibodies (TheraSAbDab) with 1,198 entries, accounting for 0.6%. * Scientific Literature (Other) with 28,735 entries, accounting for 16.9%.