Vecura

We are excited to announce that Dr. Devany West, Software Scientist at the Open Molecular Software Foundation (OMSF), will join our August 27 webinar to present on OpenADMET CheMeleon and answer your questions live. Dr. West’s work focuses heavily on the rigorous evaluation of foundation models for property prediction tasks within the ARPA-H open science initiative. She will share the story behind CheMeleon's architecture and how to execute multi-parameter ADMET profiling at scale. Expect:  🔹 A technical deep-dive into pre-trained molecular representations 🔹 Real‑world benchmark case studies on hit triaging and lead ranking 🔹 Live Q&A with the audience 📌 Register here: https://lnkd.in/gCBukDha #MedicalAI #AI4Life #OpenADMET #DigitalHealth #OMSF #ADMET

Designing novel enzyme sequences that preserve active-site residues while adapting the scaffold for a target host remains a complex bottleneck. We’ve onboarded #EnzyGen2, a 730M-parameter generative protein foundation model. It co-designs de novo amino-acid sequences and Cα backbone structures in a single autoregressive pass, conditioned on motif residues and host organism taxonomy. This joint sequence-structure generation accelerates enzyme family adaptation and protein modeling workflows, ensuring active-site geometry is strictly conserved. Vecura provides a secure, no-code Agentic AI platform, enabling research teams to test and deploy #EnzyGen2 without complex infrastructure setup. Explore the model in the Vecura blog: https://lnkd.in/gnpqhV9e Model developed by: Zhenqiao Song, Huichong Liu, Yunlong Zhao, Yang Yang, Lei Li, authors of EnzyGen2. #Vecura #ProteinModeling #AIForScience #DrugDiscovery #AgenticAI

⏰‼️Applications for AI4Life Bootcamp are now in the final week. Applications close on 23 Aug 2026. The free programme combines three live online training days with browser based practical exercises and two weeks of guided mentorship. Participants will work through target preparation, AI assisted screening, interaction and property assessment, confidence evaluation and result documentation. No programming experience or local software installation is required. 👉Wait no more! Apply now: https://lnkd.in/ghV7xFCF #AI4Life #Bootcamp #AIforScience #ComputationalBiology #DrugDiscovery #Bioinformatics #LifeSciences #TechBio #InSilico

🔎 New in Vecura Agent: choose the databases to search Before asking a question, you can select the databases most relevant to your research. This helps Vecura Agent reduce unnecessary scanning and return more focused, relevant answers faster. Not sure which databases to select? Choose all databases and let the Agent search broadly. If the selected databases do not provide enough information, Vecura Agent will recommend additional relevant databases to include. 👉 Swipe through the carousel to see how it works. 👉 Explore Vecura: https://vecura.com #Vecura #AIforScience #DrugDiscovery #Bioinformatics #ComputationalBiology #VecuraAgent

hERG channel blockade is a leading cause of cardiotoxicity-driven attrition in drug discovery. Ligand-based models often miss the structural interactions that drive inhibition. #HERGAI, now on Vecura, takes a structure-based approach: it docks compounds against the hERG receptor, extracts PLEC protein–ligand fingerprints, and predicts inhibition probability using a stacking ensemble of RF, XGBoost, and DNN classifiers. Published in J. Cheminformatics (2025), it outperformed ligand-based fingerprint baselines. This enables automated cardiotoxicity triage within agentic discovery pipelines, flagging risky compounds before costly assays. Vecura lets research teams deploy #HERGAI with no infrastructure setup. Explore the model on the Vecura blog: https://lnkd.in/gGTW7H8x Many thanks to: Viet-Khoa Tran-Nguyen, Ulrick Fineddie Randriharimanamizara, Olivier Taboureau, authors of HERGAI. #Vecura #DrugDiscovery #ADMET #AIForScience #MolecularDiscovery

⏰Only 3 days left until AI4Life Innovation Talks Webinar #3! Join host Quynh Nguyen (Director of Ecosystem Development at NYB.AI) and featured speaker Dr. Tran Minh Quan (Senior Developer Relations Manager at NVIDIA) for a technical session on Building Ambient Healthcare Agents on Modern Medical AI Infrastructure. What we'll cover: 🔹 Architecture choices for moving medical AI from demo to enterprise deployment  🔹 Speech recognition, reasoning models, and agent orchestration  🔹 Real-world applications: automated SOAP notes, patient intake, and scheduling ✅Already registered? Check your inbox for the event details! 🤔👉Not registered yet? Save your seat here: #MedicalAI #AI4Life #DigitalHealth #HealthTech

AlphaFold predictions for multi-domain proteins often contain long disordered loops that inflate structure size and complicate downstream workflows like crystallization. #AlphaCutter is now available on Vecura. It processes AlphaFold PDBs to strip non-globular regions, returning cleaned, compact globular domain fragments alongside per-residue annotations and FASTA sequences. This targeted post-processing streamlines protein modeling, ensuring only high-confidence domains advance to recombinant expression or docking. Vecura is a secure, no-code Agentic AI platform, letting research teams deploy #AlphaCutter instantly without managing DSSP dependencies or infrastructure. Explore the model in the Vecura blog: https://lnkd.in/gHMQm2zJ Many thanks to: Chunlai Tam, Wataru Iwasaki, authors of AlphaCutter. #Vecura #AIForScience #ProteinModeling #DrugDiscovery #ComputationalBiology

Structure-based drug design requires generating chemically valid, 3D-plausible small molecules that precisely fit a target protein pocket. We’ve onboarded #DrugFlow, an ICLR 2025 model integrating continuous flow matching for 3D coordinates with discrete Markov bridges for atom and bond types. It generates candidate ligands directly from a protein PDB and reference SDF. With variants like FlexFlow for joint side-chain sampling, it accelerates first-pass hit generation and molecular discovery workflows. Vecura provides a secure, no-code Agentic AI platform, letting research teams test and deploy #DrugFlow without complex infrastructure setup. Explore the model in the Vecura blog: https://lnkd.in/gPqHhmhe Many thanks to: Arne Schneuing, Ilia Igashov, Adrian Dobbelstein, Thomas Castiglione, Michael Bronstein, Bruno Correia, authors of DrugFlow. #Vecura #AIForScience #DrugDiscovery #MolecularDiscovery #AgenticAI #ComputationalBiology