parSEQ
exploratoryplatform · high · Tue Dec 12 2023 00:00:00 GMT+0000 (Coordinated Universal Time)
Improve capture of protein sequence-function datasets for protein engineering workflows, especially for AI and machine learning use cases.
Probe and Rescue Sequencing platform using a sequence-first-screen-later workflow with next-generation sequencing to pair screened variant functional data with variant sequence data.
2023 bioRxiv preprint described the parSEQ methodology and multiple protein engineering case studies.
Reported case studies include high-throughput variant DNA-template retrieval, expression-ready bacterial clone isolation, sourcing DNA for de novo designed proteins, and preparing targeted mutational libraries.
ProteinFlow
exploratoryplatform · high · Mon Sep 25 2023 00:00:00 GMT+0000 (Coordinated Universal Time)
Standardize preprocessing of protein sequence and structural data for deep learning applications in protein design, protein folding, and protein-protein interaction prediction.
Configurable Python library and computational pipeline for extracting protein organization levels from PDB-derived structure data, creating train-validation splits, and producing benchmark datasets.
2023 bioRxiv preprint presented the open-source ProteinFlow library and benchmarked a state-of-the-art protein sequence design model.
The work reports a configurable preprocessing pipeline and a curated feature-rich benchmarking dataset based on the latest annual PDB release.
Adaptyv cloud lab for protein validation
undisclosedplatform · high
Enable protein designers to upload protein sequences, order validation experiments, manage workflows, and receive experimental data for protein design decisions.
Cloud lab platform with binding, expression, and thermostability workflows, experiment management, result analysis, API access, and per-protein pricing.
Platform is described as available for ordering and managing protein validation experiments with data returned in under three weeks.
Antibody developability landscape cartography
exploratoryresearch program · medium
Quantify the plasticity of native and human-engineered antibody developability landscapes to support multi-parameter therapeutic monoclonal antibody design.
Computational analysis of 40 sequence-based and 46 structure-based developability parameters across more than two million native and human-engineered single-chain antibody sequences.
2024 Communications Biology publication reported comparative mapping of natural and engineered antibody developability landscapes.
The study found lower redundancy among structure-based developability parameters, variable sensitivity to substitutions and conformational ensembles, greater predictability for sequence-based parameters, and localization of human-engineered antibodies within subspaces of natural antibody landscapes.
