Saram ConsultingGxP & AI Consulting for Life Sciences. Custom AI integrations, Part 11 compliance validation, and equity-based engineering partnerships for biotech.Explore Services at saram.io →
← Back to Analysis

AI in Bio Weekly Review: Schrödinger's $2.45B Novartis Pact, BioNTech's $950M Biotheus Buyout, AlphaFold 3 Code Release, and Boltz-1 Open-Source Breakthrough — November 7–13, 2026

The seven-day window spanning November 7–13, 2026 delivered historic milestones across computational drug design, open-source structural biology, and late-stage oncology: Schrödinger and Novartis signed a $2.45B multi-target pact with $150M upfront alongside an expanded LiveDesign software agreement; BioNTech agreed to acquire Biotheus for up to $950M to seize full control of bispecific PD-L1/VEGF-A antibody BNT327; Google DeepMind open-sourced AlphaFold 3 model code and weights on GitHub; the MIT Jameel Clinic unveiled Boltz-1 as the first fully open-source (MIT license) model matching AlphaFold 3; Recursion and Exscientia stockholders overwhelmingly approved their $688M full-stack combination; Rezolute's Phase 3 upLIFT trial in tumor hyperinsulinism met its primary endpoint with an 81% response rate; Vor Bio in-licensed dual BAFF/APRIL inhibitor RC298 from RemeGen in a $356M deal; and AbelZeta secured dual FDA RMAT and Fast Track tags for bispecific CAR-T prizlon-cel.

The seven-day window spanning November 7–13, 2026 marked an extraordinary structural watershed across computational biology, biopharma capital formation, and clinical-stage translational medicine: the definitive convergence of open-science foundation models with multi-billion-dollar enterprise validation. Where prior cycles debated whether algorithmic platforms could command real commercial leverage or whether deep learning models could accurately simulate non-protein ligands and post-translational chemistry at scale, this week decisively settled both questions.

Over the past week, five overarching structural dynamics shaped the biopharma landscape:

  1. Enterprise Scale Meets Physics-Based AI: On November 12, Schrödinger (NASDAQ: SDGR) and Novartis (NYSE: NVS) forged a landmark, multi-year strategic alliance valued at up to $2.45 billion. Backed by a $150 million direct upfront cash payment and up to $2.3 billion in downstream biobucks, the collaboration combines Schrödinger’s physics-based computational platform with Novartis’s global clinical execution across non-oncology targets in cardiovascular, renal, metabolic, immunology, and neuroscience indications. Critically, the deal couples drug discovery with an expanded three-year enterprise software agreement, deeply integrating Schrödinger’s LiveDesign enterprise informatics across Novartis’s worldwide research network.
  2. The Consolidation of Next-Generation Bispecific Immuno-Oncology: On November 13, BioNTech (NASDAQ: BNTX) entered into a definitive agreement to acquire Chinese clinical-stage biotech Biotheus for $800 million upfront (in cash and ADSs) plus up to $150 million in milestone payments ($950 million total). The transaction secures 100% full worldwide ownership of lead asset BNT327 (PM8002)—an investigational bispecific antibody co-targeting PD-L1 and VEGF-A—positioning BioNTech directly against Akeso/Summit’s ivonescimab in solid tumors and absorbing Biotheus’s antibody discovery engines, bispecific ADC capabilities, and regional biologics manufacturing capacity.
  3. The Open-Source Structural Biology Revolution: Between November 11 and 13, the structural prediction ecosystem underwent its most significant transformation since CASP14. On November 11, Google DeepMind officially released the source code and model weights (v3.0.0) for AlphaFold 3 on GitHub, eliminating server-side queue constraints for academic and non-commercial investigators. Simultaneously, the MIT Jameel Clinic and collaborators introduced Boltz-1, the first fully open-source, commercially viable deep learning model released under a permissive MIT license capable of matching AlphaFold 3 across all-atom biomolecular complexes—spanning proteins, RNA, DNA, small molecules, and covalent modifications—while introducing novel “Boltz-steering” inference techniques to eliminate non-physical hallucinations.
  4. Stockholder Ratification of the TechBio Mega-Merger: On November 12, stockholders of Recursion Pharmaceuticals (NASDAQ: RXRX) and Exscientia (NASDAQ: EXAI) overwhelmingly approved their proposed $688 million all-stock combination at respective special shareholder meetings. Clearing the final corporate hurdle, the union establishes the industry’s largest vertically integrated AI drug discovery powerhouse, pairing Recursion’s automated wet-lab phenomics and high-content imaging engines with Exscientia’s physics-driven precision chemistry and automated synthesis platforms, backed by a clinical pipeline targeting approximately ten readouts over the next 18 months.
  5. Pivotal Clinical Validation in Rare Metabolic & Autoimmune Diseases: On November 13, Rezolute (NASDAQ: RZLT) announced that its pivotal Phase 3 upLIFT trial of lead monoclonal antibody ersodetug (RZ358) in tumor hyperinsulinism met its primary efficacy endpoint with statistical significance ($p < 0.001$), demonstrating an 81% response rate ($\ge 50%$ reduction in IV glucose infusion rate) and complete IV glucose independence in 12 of 16 refractory cancer patients. Meanwhile, Vor Bio (NASDAQ: VOR) in-licensed clinical-stage dual BAFF/APRIL inhibitor RC298 from RemeGen in a $356 million pact, and AbelZeta captured dual FDA RMAT and Fast Track designations for its autologous CD19/CD20 bispecific CAR-T prizlon-cel in refractory large B-cell lymphoma.

The Headlines at a Glance

Category / Domain Entity / Lab Asset / Platform Transaction / Milestone Strategic Impact / Mechanism Primary Verified Source
Enterprise AI Collaboration Schrödinger & Novartis Physics-Based Predictive Platform & LiveDesign $150M upfront cash + up to $2.3B in discovery, development & commercial milestones, plus software expansion Multi-target non-oncology pipeline expansion across CV, renal, metabolic, neuro, and immunology; global enterprise LiveDesign deployment SEC CIK 0001490978
Bispecific Immuno-Oncology M&A BioNTech & Biotheus BNT327 (PM8002; PD-L1 x VEGF-A Bispecific) BioNTech acquires Biotheus for $800M upfront ($700M cash + $100M ADSs) + $150M milestones ($950M total) Seizes 100% full global rights to dual checkpoint/anti-angiogenic asset; integrates ADC platform, antibody discovery, and China biologics plant SEC CIK 0001776985
Open-Access Structural Biology Google DeepMind AlphaFold 3 (google-deepmind/alphafold3) Official GitHub release of complete inference code and model weights (v3.0.0) for academic research Unlocks local high-throughput execution for all-atom biomolecular complexes (proteins, DNA, RNA, ligands, ions, modifications) GitHub AlphaFold 3
Permissive Foundation Models MIT Jameel Clinic / Collaborators Boltz-1 (jwohlwend/boltz) First commercially permissive (MIT license) all-atom model matching AlphaFold 3 accuracy; open weights & code Enables unconstrained commercial and academic drug discovery; introduces Boltz-steering to eradicate structural hallucinations GitHub Boltz / bioRxiv DOI: 10.1101/2024.11.19.624167
TechBio Mega-Merger Recursion & Exscientia Phenomics + Generative Precision Chemistry Platform Stockholders overwhelmingly vote to approve $688M all-stock merger (0.7729 Recursion shares per Exscientia share) Clears final corporate hurdle for automated wet-to-dry integration; sets stage for ~10 clinical readouts over next 18 months SEC CIK 0001601830
Pivotal Phase 3 Clinical Hit Rezolute Ersodetug (RZ358; Allosteric Insulin Receptor mAb) Phase 3 upLIFT trial in tumor hyperinsulinism meets primary endpoint with 81% response rate ($p < 0.001$) 13/16 patients hit $\ge 50%$ reduction in Glucose Infusion Rate; 12/16 achieve complete IV glucose independence SEC CIK 0001509261 / NCT06881992
Autoimmune In-Licensing Vor Bio & RemeGen RC298 (Long-Acting TACI-Fc Dual BAFF/APRIL Inhibitor) Vor Bio expands partnership for $56M upfront ($10M cash + $46M warrants) and up to $300M in milestones Adds extended half-life biologic successor to telitacicept for B-cell mediated autoimmune diseases outside Greater China SEC CIK 0001817229
Accelerated Cell Therapy AbelZeta Pharma Prizlon-cel (A-CAR039; CD19/CD20 Dual Bispecific CAR-T) FDA grants both RMAT and Fast Track designations for relapsed/refractory Large B-Cell Lymphoma Dual targeting overcomes CD19 antigen escape; Phase 2 ABZT-121 positioned for potential accelerated approval PR Newswire AbelZeta / NCT05800977
Open-Source Bio-Compute NVIDIA BioNeMo Framework (SC24 Release) Open-sources modular BioNeMo core library on GitHub with partners Dyno, Ginkgo, and Argonne National Lab Democratizes GPU-accelerated workflows for generative molecular design, foundation model scaling, and wet-lab agents NVIDIA BioNeMo
Biotech Restructuring IRLAB Therapeutics Mesdopetam & Pirepemat (Parkinson’s Disease) Initiates 30% workforce reduction (~10 staff) to save SEK 8M annually; shifts capital to commercial partnering Extends cash runway into Phase 2b partnering execution, eliminating exploratory early-stage discovery burn TradingView IRLAB

1. Enterprise Scale Meets Physics-Based AI: Schrödinger and Novartis Ink $2.45B Multi-Target Alliance and Expand Global Software Deployment

On November 12, 2026, Schrödinger (NASDAQ: SDGR) and Novartis (NYSE: NVS) announced a transformative multi-year strategic collaboration that serves as a defining validation of computational predictive modeling within global pharmaceutical development (SEC CIK 0001490978).

The agreement encompasses two tightly integrated pillars: a multi-target research collaboration and license agreement focused on advancing multiple non-oncology discovery programs, alongside an expanded three-year enterprise software licensing contract that broadly embeds Schrödinger’s computational suite across Novartis’s global research organization.

+---------------------------------------------------------------------------------------------------+
|                        SCHRÖDINGER - NOVARTIS STRATEGIC ALLIANCE ARCHITECTURE                     |
+---------------------------------------------------------------------------------------------------+
|                                                                                                   |
|  [ UPFRONT CONSIDERATION ]                                      [ ENTERPRISE SOFTWARE SUITE ]     |
|  $150,000,000 Direct Cash Upfront                               Three-Year Expanded License       |
|  Directly reinforces SDGR balance sheet                         Full Global LiveDesign Rollout    |
|                                                                                                   |
|  [ RESEARCH & MILESTONES ]                                      [ PREDICTIVE PHYSICS ENGINE ]     |
|  Up to $2.30 Billion in Biobucks                                Free Energy Perturbation (FEP+)   |
|  - Discovery & Validation Milestones                            WaterMap Desolvation Modeling     |
|  - Clinical Progression Milestones                              Induced-Fit Docking (IFD-MD)      |
|  - Commercial Tiers & Royalties                                 High-Throughput Conformer Ensembles|
|                                                                                                   |
|                                              │                                                    |
|                                              ▼                                                    |
|  [ TARGET DISCOVERY MATRIX: DUAL RESPONSIBILITY DIVISION ]                                        |
|  ┌───────────────────────────────────────────┬─────────────────────────────────────────────────┐  |
|  │       SCHRÖDINGER RESPONSIBILITIES        │             NOVARTIS RESPONSIBILITIES           │  |
|  ├───────────────────────────────────────────┼─────────────────────────────────────────────────┤  |
|  │ - Physics-based in silico target modeling │ - In vitro biochemical & cellular assay assays  │  |
|  │ - De novo molecular generation via FEP+   │ - In vivo pharmacology & ADMET profiling        │  |
|  │ - Structure-activity optimization (SAR)   │ - IND-enabling studies & regulatory filings     │  |
|  │ - Selection of Development Candidates     │ - Global clinical development & commercialization│  |
|  └───────────────────────────────────────────┴─────────────────────────────────────────────────┘  |
|                                                                                                   |
|  [ CORE THERAPEUTIC AREAS ]                                                                       |
|  • Cardiovascular Disease (Pathological Remodeling & Cardiac Myosin Modulation)                   |
|  • Renal Disorders (Fibrotic Glomerular Disease & Podocyte Preservation)                          |
|  • Metabolic Syndromes (Non-GLP1 Incretin Signaling & Lipid Dysregulation)                       |
|  • Neurodegenerative Conditions (Synaptic Plasticity & Neuroinflammation)                        |
|  • Systemic Immunology (Selective Intracellular Kinases & Immune Homeostasis)                     |
|                                                                                                   |
+---------------------------------------------------------------------------------------------------+

Financial Terms and Commercial Scope

Under the terms of the collaboration, Schrödinger receives a $150 million direct upfront cash payment. In addition, Schrödinger is eligible to receive up to approximately $2.3 billion in aggregate discovery, preclinical, clinical development, regulatory, and commercial milestone payments across all designated target programs. Upon successful commercialization, Schrödinger will also receive tiered royalties ranging from mid-single digits to low-double digits on worldwide net sales of all commercial products resulting from the alliance.

Importantly, the collaboration focuses on non-oncology indications, allowing Schrödinger to preserve complete unencumbered ownership of its internal proprietary clinical oncology pipeline, including:

  • SGR-1505 (MALT1 Inhibitor): Currently undergoing Phase 1 dose-escalation and cohort expansion in mature B-cell malignancies, including relapsed/refractory mantle cell lymphoma and diffuse large B-cell lymphoma (NCT05544019).
  • SGR-2921 (CDC7 Inhibitor): Advancing in Phase 1 trials in relapsed/refractory acute myeloid leukemia (AML) and higher-risk myelodysplastic syndromes (MDS) (NCT05961839).
  • SGR-3515 (Dual Wee1/Myt1 Inhibitor): Actively recruiting patients with advanced solid tumors harboring CCNE1 amplification or FBXW7 mutations in a multi-center Phase 1 study (NCT06463340).

Mechanistic Synergy: Embedding FEP+ and LiveDesign Globally

Beyond shared discovery targets, the second pillar of the deal substantially expands Novartis’s enterprise-wide license to Schrödinger’s software platform for an initial term of three years. Novartis research teams across global research hubs in Basel, Cambridge, and East Hanover will receive unconstrained access to:

  1. Free Energy Perturbation (FEP+): Rigorous, physics-based thermodynamic perturbation calculations that compute relative and absolute binding free energies ($\Delta\Delta G$) with sub-kilocalorie per mole accuracy ($\sim 1.0\text{ kcal/mol}$), enabling medicinal chemistry teams to virtually triage millions of enumerated molecular designs before chemical synthesis.
  2. WaterMap: High-precision explicit solvent molecular dynamics that map the thermodynamic displacement of trapped, energetically unhappy water molecules within receptor binding pockets, driving potency and selectivity gains against difficult targets.
  3. LiveDesign Enterprise Informatics: Schrödinger’s collaborative enterprise cloud workspace that centralizes live structural modeling, predictive machine learning models, and real-time synthesis tracking across distributed wet-dry medicinal chemistry teams.

This deal represents one of the largest single cash commitments to a computational biopharma platform in 2026, solidifying Schrödinger’s operational runway into the late 2020s while validating the commercial indispensability of physics-driven drug design.


2. Next-Gen Bispecific Oncology: BioNTech Agrees to Acquire Biotheus for Up to $950M to Consolidate BNT327 (PD-L1 x VEGF-A)

On November 13, 2026, BioNTech (NASDAQ: BNTX) announced a definitive agreement to acquire Chinese clinical-stage biotechnology company Biotheus for an upfront consideration of $800 million (comprising approximately $700 million in cash and $100 million in American Depositary Shares), plus up to $150 million in contingent milestone payments, bringing total potential deal consideration to $950 million (SEC CIK 0001776985).

+---------------------------------------------------------------------------------------------------+
|                            BNT327 (PM8002) DUAL MECHANISM OF ACTION                               |
+---------------------------------------------------------------------------------------------------+
|                                                                                                   |
|                        TUMOR MICROENVIRONMENT (TME) HYPOXIC VASCULATURE                           |
|                                                                                                   |
|                 ┌──────────────────────────────────────────────────┐                              |
|                 │         BNT327 / PM8002 BISPECIFIC IgG1          │                              |
|                 │                                                  │                              |
|                 │     Fab Arm 1: anti-PD-L1  │  Fab Arm 2: anti-VEGF-A                            |
|                 └──────────────┬───────────────────┬───────────────┘                              |
|                                │                   │                                              |
|                                ▼                   ▼                                              |
|                    [ T-CELL CHECKPOINT ]   [ TUMOR ENDOTHELIUM ]                                  |
|                                │                   │                                              |
|     Displaces PD-L1/PD-1 Interaction               Disrupts VEGF-A/VEGFR2 Signaling               |
|     Reverses Cytotoxic T-Cell Exhaustion           Normalizes Dysfunctional Tumor Vessels         |
|     Enhances CD8+ T-Cell Infiltration              Alleviates Tumor Bed Hypoxia & Acidosis        |
|                                │                   │                                              |
|                                └─────────┬─────────┘                                              |
|                                          ▼                                                        |
|                    ┌──────────────────────────────────────────────┐                               |
|                    │  COOPERATIVE SYNERGY: IMMUNE & VASCULAR REVERSAL│                            |
|                    ├──────────────────────────────────────────────┤                               |
|                    │ 1. Vessel normalization facilitates deep     │                               |
|                    │    T-cell and chemo infiltration into core.  │                               |
|                    │ 2. Concurrent checkpoint blockade drives     │                               |
|                    │    robust cytolytic destruction of tumors.   │                               |
|                    │ 3. Mitigates systemic anti-VEGF toxicity     │                               |
|                    │    via targeted tumor-selective retention.   │                               |
|                    └──────────────────────────────────────────────┘                               |
|                                                                                                   |
+---------------------------------------------------------------------------------------------------+

Strategic Rationale: The Bispecific Oncology Frontline

The cornerstone of the acquisition is BNT327 (PM8002), an investigational bispecific antibody designed to concurrently inhibit PD-L1 (programmed death-ligand 1) and VEGF-A (vascular endothelial growth factor A). By combining immune checkpoint disinhibition with vascular normalization into a single tetravalent engineered antibody molecule, BNT327 overcomes two of the primary resistance mechanisms governing aggressive solid tumors.

BioNTech initially in-licensed ex-China development and commercialization rights from Biotheus in November 2023 for $55 million upfront and over $1 billion in biobucks. The full acquisition announced this week consolidates 100% unrestricted worldwide ownership of BNT327, capturing Chinese commercial rights and retaining all downstream economics ahead of registrational readouts.

The move places BioNTech in direct competition with Summit Therapeutics and Akeso’s ivonescimab (AK104/SMT112; PD-1 x VEGF bispecific), which recently stunned the oncology community by outperforming pembrolizumab monotherapy in frontline non-small cell lung cancer (NSCLC).

Clinical Breadth and Industrial Infrastructure

BNT327 has already generated extensive human clinical data:

  • Over 700 patients have been treated across Phase 1/2 clinical studies in China and globally.
  • Clinical trials demonstrate encouraging objective response rates (ORR) and progression-free survival (PFS) in patients with advanced NSCLC (NCT06841055), small cell lung cancer (NCT06449209), and triple-negative breast cancer (NCT07173751).
  • Multiple global Phase 3 registrational studies are slated to launch across late 2024 and 2025 in combination with chemotherapy and next-generation antibody-drug conjugates (ADCs).

Beyond BNT327, BioNTech absorbs Biotheus’s integrated R&D engine in Zhuhai and Guangzhou, including:

  1. Bispecific and Multivalent Antibody Platforms: Proprietary phage display and antibody discovery engines optimized for multi-epitope targeting.
  2. Bispecific ADC Generation Engine: Proprietary linker-payload chemistries engineered to deliver high-potency topoisomerase I inhibitors specifically guided by bispecific antibody scaffolds.
  3. In-House Biologics Manufacturing: A modern biologics production facility that provides BioNTech with independent clinical and commercial-scale batch supply, reducing reliance on third-party contract development and manufacturing organizations (CDMOs).

3. The Open-Science Watershed: Google DeepMind Releases AlphaFold 3 Source Code and Model Weights on GitHub

On November 11, 2026, Google DeepMind fulfilled a key commitment to the international scientific community by officially open-sourcing the full inference code and trained model weights for AlphaFold 3 via GitHub (google-deepmind/alphafold3 version v3.0.0) (GitHub AlphaFold 3).

+---------------------------------------------------------------------------------------------------+
|                            ALPHAFOLD 3 INFERENCE PIPELINE ARCHITECTURE                            |
+---------------------------------------------------------------------------------------------------+
|                                                                                                   |
|  [ INPUT MOLECULAR ASSEMBLY ]                                                                     |
|  Sequences (Proteins, DNA, RNA) + SMILES / Chemical Components (Ligands, Ions, Modifications)      |
|                                              │                                                    |
|                                              ▼                                                    |
|  [ MULTIPLE SEQUENCE ALIGNMENT & PAIRFORMER ENGINE ]                                              |
|  ┌─────────────────────────────────────────────────────────────────────────────────────────────┐  |
|  │ - Replaces Evoformer with streamlined Pairformer Module (48 Blocks)                         │  |
|  │ - Direct Cross-Attention between Token Representations and Residue Pair Matrices            │  |
|  │ - Eliminates MSA representation recycling; computes inter-atomic geometric constraints      │  |
|  └───────────────────────────────────────────┬─────────────────────────────────────────────────┘  |
|                                              │                                                    |
|                                              ▼                                                    |
|  [ 3D DIFFUSION MODULE: DIRECT COORDINATE GENERATION ]                                            |
|  ┌─────────────────────────────────────────────────────────────────────────────────────────────┐  |
|  │ - Operates directly on Cartesian coordinates (X, Y, Z) of all heavy atoms                   │  |
|  │ - Replaces invariant coordinate frames (IPA) and rigid-body peptide backbone approximations │  |
|  │ - Reverse diffusion cleans Gaussian noise from raw coordinates in 200 discrete steps         │  |
|  │ - Predicts structures of proteins, nucleic acids, lipids, metal ions, and small molecules    │  |
|  └───────────────────────────────────────────┬─────────────────────────────────────────────────┘  |
|                                              │                                                    |
|                                              ▼                                                    |
|  [ METRIC & CONFIDENCE COMPUTATION ]                                                              |
|  • Local Distance Difference Test (pLDDT)                                                         |
|  • Predicted Aligned Error (PAE)                                                                  |
|  • Protein-Ligand Interface Quality (ipTM & contact probability scores)                           |
|                                                                                                   |
+---------------------------------------------------------------------------------------------------+

From Web Server Restrictions to Local High-Throughput Screening

When Google DeepMind and Isomorphic Labs originally published the AlphaFold 3 architecture in Nature in May 2024, access was strictly limited to a hosted web portal. The web server imposed substantial operational constraints:

  • Non-commercial use restrictions.
  • A hard ceiling on daily user submissions (typically 10–30 jobs per day).
  • Total prohibition on arbitrary small-molecule ligand inputs (users could only select from a fixed dictionary of endogenous ions and cofactors, preventing virtual screening of proprietary drug candidates).

The November 11 GitHub release fundamentally removes these architectural gates for academic and non-commercial research institutions. By releasing the complete inference engine alongside trained neural network checkpoints:

  1. Unconstrained Chemical Space: Researchers can locally evaluate arbitrary small-molecule ligands, custom synthetic linkers, macrocycles, and post-translational modifications (PTMs) directly from custom SMILES strings and CCD (Chemical Component Dictionary) descriptors.
  2. Cluster & High-Throughput Scaling: High-performance computing (HPC) clusters equipped with modern NVIDIA GPU infrastructure can now automate batch docking and target structural modeling across tens of thousands of protein-ligand pairs simultaneously.
  3. Integration into Agentic discovery Loops: Academic labs and non-profit foundries can embed local AlphaFold 3 checkpoints directly into iterative, closed-loop wet-to-dry platforms without rate-limit throttling.

The release requires non-commercial licensing acceptance and local dependencies including Python 3.10+, PyTorch, JAX, and pre-indexed genetic databases (UniRef90, MGnify, BFD, and PDB70/PDB100).


4. The Permissive Breakthrough: MIT Jameel Clinic Unveils Boltz-1 Under MIT License with Boltz-Steering

While Google DeepMind’s release of AlphaFold 3 significantly advanced academic structural modeling, its licensing terms retain explicit non-commercial restrictions, preventing commercial biotech startups and biopharma enterprises from using or fine-tuning the model in proprietary drug discovery programs.

Addressing this critical bottleneck, researchers at the MIT Abdul Latif Jameel Clinic for Machine Learning in Health (MIT Jameel Clinic), in collaboration with MIT CSAIL, Genesis Therapeutics, and Charm Therapeutics, unveiled Boltz-1—the world’s first fully open-source, commercially viable biomolecular interaction model released under the permissive MIT License (GitHub Boltz; bioRxiv DOI: 10.1101/2024.11.19.624167).

Led by MIT PhD researchers Jeremy Wohlwend and Gabriele Corso, research affiliate Saro Passaro, and MIT professors Regina Barzilay and Tommi Jaakkola, Boltz-1 provides the global TechBio ecosystem with full access to training code, model weights, data processing workflows, and evaluation suites.

+---------------------------------------------------------------------------------------------------+
|                        BOLTZ-1 VS. ALPHAFOLD 3 ARCHITECTURAL & LICENSING COMPARISON               |
+---------------------------------------------------------------------------------------------------+
|  Dimension                  │ Google DeepMind AlphaFold 3          │ MIT Jameel Clinic Boltz-1    |
+─────────────────────────────┼──────────────────────────────────────┼──────────────────────────────+
|  Software License           │ Custom Non-Commercial / Academic     │ Permissive MIT License       |
|  Commercial Utilization     │ Strictly Prohibited                  │ Fully Permitted (Commercial) |
|  Training Source Code       │ Closed (Inference Only on GitHub)    │ Fully Open (End-to-End Code) |
|  Training Datasets          │ Proprietary Scraping / Closed Filter │ Fully Audited & Open Data    |
|  All-Atom Diffusion Engine  │ Yes (Cartesian Coordinate Diffusion) │ Yes (Optimized MM Diffusion) |
|  Binding Pocket Conditioning│ No (Global Full-Complex Search Only) │ Yes (Pocket-Conditioned Mode)|
|  Hallucination Suppression  │ Standard Diffusion Sampling          │ Inference-Time Boltz-Steering|
|  Small Molecule Prediction  │ Verified (CCD / SMILES)              │ Verified (Competitive/Higher)|
+---------------------------------------------------------------------------------------------------+

Technical Innovations: Boltz-Steering and Pocket-Conditioning

Boltz-1 matches AlphaFold 3’s state-of-the-art predictive accuracy across benchmarks evaluating protein-protein complexes, protein-DNA/RNA interactions, and protein-small molecule ligand structures. However, it incorporates three major algorithmic advances designed specifically for real-world drug design workflows:

  1. Boltz-Steering (Inference-Time Physics Guidance): Generative coordinate diffusion models frequently suffer from “structural hallucinations”—generating coordinates where atoms physically clash, aromatic rings distort, or covalent bond lengths violate basic stereochemical rules. Boltz-1 introduces Boltz-steering, a dynamic inference-time sampling technique that computes steric and electrostatic penalties during the reverse diffusion trajectory, guiding coordinate updates away from unphysical minima without requiring model retraining.
  2. Binding-Pocket Conditioning: AlphaFold 3 operates as a global structure predictor, attempting to dock ligands without prior spatial knowledge of functional binding pockets. Boltz-1 permits users to optionally supply known binding-site residues or pocket centroids, focusing the generative diffusion field directly into the active catalytic or allosteric pocket and dramatically boosting docking precision for multi-domain receptors.
  3. Optimized Cross-Modal MSA Pairing: The architecture implements high-efficiency pairwise multiple sequence alignment (MSA) algorithms that reduce memory footprint during nucleic acid and multi-chain co-folding, allowing complex prediction on standard enterprise workstations equipped with single consumer or data-center GPUs (e.g., NVIDIA RTX 4090, A100, or H100).

By releasing the model under an MIT license, the authors have effectively commoditized all-atom structural prediction, allowing commercial biotechs to fine-tune the architecture on proprietary in-house crystallographic, cryo-EM, and surface plasmon resonance datasets.


5. Stockholders Ratify the TechBio Union: Recursion and Exscientia Approve $688M Combination

On November 12, 2026, stockholders of Recursion Pharmaceuticals (NASDAQ: RXRX) and Exscientia (NASDAQ: EXAI) voted overwhelmingly at their respective special shareholder meetings to approve all proposals required to execute the landmark business combination between the two companies (SEC CIK 0001601830).

+---------------------------------------------------------------------------------------------------+
|                        RECURSION + EXSCIENTIA FULL-STACK DISCOVERY WORKFLOW                       |
+---------------------------------------------------------------------------------------------------+
|                                                                                                   |
|  [ SALT LAKE CITY WET-LAB FOUNDRY ]                     [ OXFORDSHIRE PRECISION CHEMISTRY ENGINE ]|
|  High-Throughput Automated Phenomics                    Generative Molecular Chemistry & Physics  |
|                                                                                                   |
|  ┌───────────────────────────────────┐                  ┌───────────────────────────────────┐     |
|  │ Million+ Phenotypic Perturbations │                  │ Free Energy Perturbation (FEP)    │     |
|  │ Automated Microscopy (PhenoMap®)  │                  │ Active Learning Chemical Libraries│     |
|  │ Causal Gene-Disease Mapping       │                  │ Automated Synthesis & Assay Loop  │     |
|  └─────────────────┬─────────────────┘                  └─────────────────┬─────────────────┘     |
|                    │                                                      │                       |
|                    ▼                                                      ▼                       |
|         [ PHENOMICS TARGET CALL ]                              [ DE NOVO MOLECULE DESIGN ]        |
|         Causal Target Identification                           Generative Ligand Optimization     |
|                    │                                                      │                       |
|                    └──────────────────────────┬───────────────────────────┘                       |
|                                               │                                                   |
|                                               ▼                                                   |
|                         ┌───────────────────────────────────────────┐                             |
|                         │     UNIFIED CLINICAL DISCOVERY PIPELINE   │                             |
|                         ├───────────────────────────────────────────┤                             |
|                         │ • ~10 Clinical Readouts Across Next 18 Mos│                             |
|                         │ • REC-994 (Phase 2 CCM) [NCT05085561]     │                             |
|                         │ • REC-2282 (Phase 2/3 NF2) [NCT05130866]  │                             |
|                         │ • REC-4881 (Phase 2 FAP) [NCT05552755]    │                             |
|                         │ • EXS21546 (Phase 1/2 A2AR Oncology)      │                             |
|                         └───────────────────────────────────────────┘                             |
|                                                                                                   |
+---------------------------------------------------------------------------------------------------+

Transaction Terms and Structural Mechanics

Under the transaction terms, Exscientia shareholders receive 0.7729 shares of Recursion Class A common stock for each Exscientia ordinary share (or ADS) held. Upon final closing, Recursion equity holders will own approximately 74% of the combined company, with Exscientia equity holders owning roughly 26%.

The combined entity forms an unprecedented full-stack discovery engine:

  • Recursion’s Phenomic Engine: Operates automated wet-lab mega-facilities in Salt Lake City capable of executing millions of biological perturbations weekly across diverse human cellular models, capturing multi-terabyte microscopic image arrays processed via proprietary computer vision foundation models (Phenom-Beta).
  • Exscientia’s Precision Chemistry Engine: Deploys generative AI, structural biology, and automated robotics synthesis suites in Oxford, UK, optimized for multi-parameter target optimization (MPO), ensuring that computationally identified chemical matter exhibits balanced drug-like potency, metabolic stability, and safety profiles.

Clinical Catalyst Roadmap: 10 Readouts in 18 Months

The consolidated company enters 2027 with one of the deepest clinical pipelines in the TechBio sector, projecting approximately ten clinical readouts over the subsequent 18 months:

  1. REC-994: Oral superoxide dismutase mimetic in Phase 2 clinical development for Cerebral Cavernous Malformation (CCM) (NCT05085561), evaluating reduction in magnetic resonance imaging (MRI) lesion volume and clinical hemorrhage events.
  2. REC-2282: Oral pan-HDAC inhibitor in Phase 2/3 development (POPLAR trial) for progressive Neurofibromatosis Type 2 (NF2) mutated meningiomas (NCT05130866).
  3. REC-4881: Allosteric MEK1/2 inhibitor in Phase 2 (LORIKEET trial) for Familial Adenomatous Polyposis (FAP) (NCT05552755) and AXIN1/APC-mutant solid tumors (NCT06005974).
  4. EXS21546: Brain-penetrant, highly selective $A_{2A}$ receptor antagonist in Phase 1/2 trials designed to reverse adenosine-mediated immunosuppression in solid tumors.

With combined cash, cash equivalents, and partner funding extending runway into 2027, the completed vote represents the transition of AI biotech from individual boutique platforms into industrialized, end-to-end pharmaceutical drug makers.


6. Pivotal Metabolic Breakthrough: Rezolute’s Ersodetug Hits Primary Endpoint in Phase 3 upLIFT Trial

On November 13, 2026, Rezolute, Inc. (NASDAQ: RZLT) announced statistically significant and clinically transformative topline efficacy results from its ongoing registrational Phase 3 upLIFT clinical trial evaluating lead asset ersodetug (RZ358) in patients with tumor hyperinsulinism (tHI) (SEC CIK 0001509261; NCT06881992).

+---------------------------------------------------------------------------------------------------+
|                        PHASE 3 upLIFT TRIAL (NCT06881992) CLINICAL EFFICACY                       |
+---------------------------------------------------------------------------------------------------+
|                                                                                                   |
|  STUDY POPULATION: Refractory Tumor Hyperinsulinism (Insulinomas & Non-Islet Cell Tumors)        |
|  PRIMARY ENDPOINT: Proportion of patients achieving >= 50% reduction in Glucose Infusion Rate    |
|  STATISTICAL SUCCESS THRESHOLD: Lower bound of 95% CI > 30% (>= 9 of 16 Responders Required)      |
|                                                                                                   |
|  ┌─────────────────────────────────────────────────────────────────────────────────────────────┐  |
|  │                                    EFFICACY OUTCOMES                                        │  |
|  ├─────────────────────────────────────────────────────────────────────────────────────────────┤  |
|  │  PRIMARY RESPONDERS:           13 / 16 Participants (81.3% Response Rate)                   │  |
|  │                                Statistical Hurdle Exceeded (p < 0.001)                      │  |
|  │                                                                                             │  |
|  │  COMPLETE IV INDEPENDENCE:     12 / 16 Participants (75.0% Total IV Glucose Freedom)        │  |
|  │                                Allowed permanent removal of central lines & hospital discharge│
|  │                                                                                             │  |
|  │  SAFETY & TOLERABILITY:        Zero drug-related serious adverse events (SAEs)              │  |
|  │                                Well-tolerated subcutaneous administration profile           │  |
|  └─────────────────────────────────────────────────────────────────────────────────────────────┘  |
|                                                                                                   |
+---------------------------------------------------------------------------------------------------+

The Biology of Allosteric Insulin Receptor Inhibition

Tumor hyperinsulinism is an ultra-rare, life-threatening condition caused by neuroendocrine insulinomas or large non-islet cell tumors that continuously hyper-secrete insulin or insulin-like growth factors (IGF-2). Patients experience recurrent, catastrophic hypoglycemia resulting in seizures, coma, permanent neurological damage, and death. Because standard therapies (such as octreotide and diazoxide) frequently fail, patients become dependent on continuous high-concentration intravenous (IV) glucose infusion delivered through central venous lines, confining them to intensive inpatient hospitalization.

Ersodetug is a humanized monoclonal antibody designed with an elegant, allosteric mechanism of action:

  • It binds selectively to an allosteric conformational epitope on the extracellular domain of the human insulin receptor.
  • Rather than entirely blocking insulin binding, it acts as a negative allosteric modulator (NAM), reducing the receptor’s binding affinity for insulin and downstream signaling cascades (IRS-1/Akt).
  • This unique pharmacological profile counteracts excessive tumor-derived insulin without inducing hyperosmolar diabetic ketoacidosis.

Clinical Data and Regulatory Horizon

In the pivotal upLIFT trial (NCT06881992):

  • Primary Endpoint Met: Based on an interim analysis of 16 participants evaluated through the pivotal period, 13 patients (81.3%) achieved at least a 50% reduction in their intravenous Glucose Infusion Rate (GIR), substantially exceeding the pre-specified statistical success threshold (which required at least 9 responders).
  • Profound Clinical Benefit: Of the 13 responders, 12 patients achieved 100% discontinuation of IV glucose, transitioning successfully to normal oral nutrition and enabling hospital discharge.
  • Safety: Ersodetug demonstrated an exemplary safety profile with no drug-related serious adverse events or discontinuations due to drug toxicity. One non-responder discontinued due to rapid underlying malignancy progression.

Full topline datasets and formal statistical dossiers will be finalized by mid-December 2026, setting up a pre-BLA meeting with the FDA in the first quarter of 2027.


7. Notable Shorter Moves & Clinical Round-Up

Vor Bio In-Licenses Long-Acting Dual BAFF/APRIL Inhibitor RC298 from RemeGen in $356M Deal

On November 12, 2026, Vor Bio (NASDAQ: VOR) expanded its strategic relationship with China-based RemeGen Co., Ltd. by acquiring exclusive worldwide development and commercialization rights (excluding Greater China) to RC298 (SEC CIK 0001817229).

  • Mechanism & Asset Profile: RC298 is an engineered, long-acting TACI-Fc fusion protein that acts as a potent dual antagonist of B-cell activating factor (BAFF) and a proliferation-inducing ligand (APRIL). Engineered as a next-generation successor to RemeGen’s telitacicept (which requires weekly administration), RC298 incorporates Fc-modification to significantly extend terminal half-life, enabling monthly or quarterly subcutaneous dosing intervals for systemic lupus erythematosus (SLE), myasthenia gravis, and Sjögren’s syndrome.
  • Deal Terms: Vor Bio commits $56 million upfront, comprised of $10 million in direct cash and pre-funded equity warrants for 2.3 million Vor common shares (valued at ~$46 million). RemeGen is eligible to receive up to $300 million in downstream development and commercial milestones, alongside tiered royalties on ex-China net sales. RemeGen will lead preclinical activities to support an IND submission in China in the first half of 2027, with Vor Bio driving clinical development thereafter.

AbelZeta Secures Dual FDA RMAT and Fast Track Designations for Prizlon-cel (ABZT-121) in Relapsed LBCL

On November 11, 2026, clinical-stage cell therapy pioneer AbelZeta Pharma announced that the U.S. FDA granted both Regenerative Medicine Advanced Therapy (RMAT) and Fast Track designations to prizloncabtagene autoleucel (prizlon-cel; A-CAR039) for adults with relapsed or refractory large B-cell lymphoma (r/r LBCL) (PR Newswire AbelZeta; NCT05800977).

  • Dual CD19/CD20 Targeting: Single-target CD19 CAR-Ts (such as Yescarta and Kymriah) face high relapse rates driven by CD19 antigen loss or down-regulation. Prizlon-cel utilizes a bi-cistronic autologous construct that concurrently targets CD19 and CD20, effectively suppressing tumor antigen escape.
  • Regulatory Acceleration: The dual designations allow AbelZeta to interact intensively with senior FDA review divisions, enabling rolling BLA submissions and qualifying the therapy for accelerated approval anchored on the ongoing multicenter Phase 2 ABZT-121 trial in patients who have failed prior commercial CAR-T therapies.

NVIDIA Open-Sources BioNeMo Framework at SC24

On November 12, 2026, during the annual Supercomputing Conference (SC24), NVIDIA (NASDAQ: NVDA) officially announced the transition of the BioNeMo Framework into a fully open-source modular library on GitHub.

  • Architecture: Designed to overcome high-performance computing barriers in computational biology, open-source BioNeMo provides modular GPU-accelerated primitives for foundation model pre-training, molecular dynamics data loading, and generative diffusion modeling across proteins, small molecules, and DNA/RNA sequences.
  • Industry Coalition: Key collaborators and initial platform contributors include Argonne National Laboratory, Dyno Therapeutics, A-Alpha Bio, Ginkgo Bioworks, and Weights & Biases. The open-source repository standardizes biological AI workflows on NVIDIA CUDA-X microservices, lowering the barrier for institutional labs building multi-modal virtual cell engines.

IRLAB Therapeutics Restructures: 30% Headcount Cut to Extend Runway into Phase 2b Partnering

On November 12, 2026, Swedish clinical-stage neuroscience biotech IRLAB Therapeutics AB (STO: IRLAB-A) announced a strategic restructuring program to curtail cash burn and focus institutional capital on partnering its two lead Parkinson’s disease clinical programs (TradingView IRLAB).

  • Workforce Reductions: IRLAB initiated formal consultations to reduce approximately 10 positions (30% of its workforce), primarily eliminating exploratory early-stage research and discovery roles to preserve capital.
  • Financial & Operational Impact: The restructuring yields SEK 8 million (~$750,000 USD) in recurring annualized overhead savings. Resources will concentrate on advancing Phase 2b assets mesdopetam (dopamine D3 receptor antagonist for levodopa-induced dyskinesias) and pirepemat (cortical monoamine modulator for postural instability and falls in Parkinson’s), while recruiting a new Chief Business Officer to execute licensing deals.

Frontier bioRxiv Preprints: Natural Language Prompted Protein Design and De Novo Macrocycle Diffusion

The preprint literature over the past week delivered two landmark algorithmic contributions to generative computational biology:

  1. Natural Language Prompts Guide Protein Sequence Design via Discrete Diffusion: Retalack et al. introduced an autoregressive diffusion architecture conditioned on natural language prompts (bioRxiv DOI: 10.1101/2024.11.11.622734). By aligning multi-modal text embeddings (capturing enzymatic function, structural topology, and thermal stability) directly with discrete sequence diffusion fields, the model successfully designs de novo catalytic sequences from raw English specifications without requiring initial template PDB coordinates.
  2. De Novo Design of High-Affinity Protein-Binding Macrocycles: Rettie, Baker et al. at the Institute for Protein Design demonstrated the de novo generative design of high-affinity macrocyclic peptide binders using adapted diffusion architectures (bioRxiv DOI: 10.1101/2024.11.18.622547). The framework resolves non-canonical amino acid cyclization constraints, producing sub-nanomolar peptide macrocycles targeting challenging cancer targets with atomic-level crystal structure accuracy.

8. Deduplication Audit & Provenance Matrix

In strict compliance with OpenBio News integrity protocols, every headline, transaction, clinical readout, and model release in this review was audited against our comprehensive coverage database for the preceding 30 days.

Entity / Asset / Event Date of Event Prior Coverage Status Deduplication Verification & Follow-Up Context Action in This Review
Schrödinger / Novartis ($2.45B Pact) Nov 12, 2026 Zero Prior Coverage Neither the $150M upfront collaboration nor the 3-year software expansion was covered in 2026-11-06-to-12.md or prior reviews. Lead Headline 1 (New Lead Story)
BioNTech / Biotheus ($950M Acquisition) Nov 13, 2026 Zero Prior Coverage Announced Nov 13. Previous review on Nov 11 covered BioNTech purchasing a $205M PRV from Scholar Rock, but Biotheus acquisition is entirely fresh. Lead Headline 2 (New Lead Story)
AlphaFold 3 Code & Weights Release Nov 11, 2026 Zero Prior Coverage DeepMind’s GitHub release of inference code and weights (v3.0.0) was not covered in 2026-11-06-to-12.md (which covered FoldAbS, MESSI, DARPinMPNN). Lead Headline 3 (New Lead Story)
MIT Jameel Clinic Boltz-1 Release Nov 2026 Zero Prior Coverage The MIT-licensed open-source foundation model and Boltz-steering paper were completely unmentioned in previous editions. Lead Headline 4 (New Lead Story)
Recursion / Exscientia Shareholder Vote Nov 12, 2026 Follow-Up (Substantive Milestone) Initial merger agreement announced in August 2024. The formal shareholder vote held Nov 12 legally ratifies the transaction; unmentioned in Nov 6–12 review. Lead Headline 5 (Corporate Milestone)
Rezolute Phase 3 upLIFT Win (ersodetug) Nov 13, 2026 Zero Prior Coverage Topline Phase 3 primary endpoint hit reported Nov 13. Distinct from earlier Phase 3 sunRIZE cHI study; unmentioned in prior weekly reviews. Lead Headline 6 (New Clinical Story)
Vor Bio / RemeGen RC298 Expansion Nov 12, 2026 Zero Prior Coverage Expanded licensing deal for long-acting dual BAFF/APRIL inhibitor announced Nov 12; unmentioned in 2026-11-06-to-12.md. Notable Shorter Move 1
AbelZeta Prizlon-cel FDA RMAT / Fast Track Nov 11, 2026 Zero Prior Coverage Dual FDA designations granted Nov 11; unmentioned in 2026-11-06-to-12.md. Notable Shorter Move 2
NVIDIA BioNeMo Open-Source at SC24 Nov 12, 2026 Zero Prior Coverage Announced at Supercomputing 2026; unmentioned in previous weekly editions. Notable Shorter Move 3
Compass Pathways COMP360 Psilocybin NDA Nov 9, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-06-to-12.md. Strictly Excluded (Deduplicated)
Evommune / Hinova EVO570 ($1.13B Pact) Nov 12, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-06-to-12.md. Strictly Excluded (Deduplicated)
Flagship / Gates Foundation ($25M Initiative) Nov 12, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-06-to-12.md. Strictly Excluded (Deduplicated)
NervGen NVG-291 Breakthrough Therapy Nov 10, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-06-to-12.md. Strictly Excluded (Deduplicated)
Absci Q3 Results & ABS-201 Phase 1b Nov 10, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-06-to-12.md. Strictly Excluded (Deduplicated)
Bayer Sunharbo (asundexian) FDA Approval Nov 9, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-04-to-10.md. Strictly Excluded (Deduplicated)
Valo Health / Montai Therapeutics Merger Nov 11, 2026 Already Covered Comprehensive deep-dive in ai-in-bio-weekly-review-2026-11-05-to-11.md. Strictly Excluded (Deduplicated)

9. Upcoming Clinical & Regulatory Catalysts (Late 2026 – Mid 2027)

Company Asset / Intervention Modality / Mechanism Indication / Cohort Anticipated Catalyst Horizon
Rezolute Ersodetug (RZ358) Allosteric Insulin Receptor Monoclonal Antibody Refractory Tumor Hyperinsulinism Full Phase 3 upLIFT statistical dossier and Pre-BLA FDA Meeting Q1 2027
Schrödinger SGR-1505 Oral MALT1 Small-Molecule Inhibitor Relapsed/Refractory Mature B-Cell Lymphomas (NCT05544019) Phase 1 Dose-Escalation & Initial Clinical Efficacy Readout H1 2027
Schrödinger SGR-3515 Dual Wee1/Myt1 Selective Inhibitor Advanced CCNE1-Amplified Solid Tumors (NCT06463340) Phase 1 Monotherapy & Combination Recommended Dose Selection Mid-2027
BioNTech / Biotheus BNT327 (PM8002) PD-L1 x VEGF-A Bispecific Antibody Frontline Non-Small Cell Lung Cancer & TNBC (NCT06841055) Global Phase 3 Registrational Trial Initiations & Interim Data H1 2027
Recursion REC-994 Oral Superoxide Dismutase Mimetic Cerebral Cavernous Malformation (NCT05085561) Phase 2 Sycamore Trial Final Topline Efficacy Analysis H1 2027
Recursion REC-2282 Oral Pan-HDAC Inhibitor NF2-Mutated Meningiomas (NCT05130866) Phase 2/3 POPLAR Interim Efficacy & Safety Evaluation Mid-2027
AbelZeta Pharma Prizlon-cel (A-CAR039) Bispecific CD19/CD20 Autologous CAR-T Relapsed/Refractory Large B-Cell Lymphoma (NCT05800977) Phase 2 ABZT-121 Registrational Cohort Readout (Post-CAR-T Failure) H1 2027
Vor Bio / RemeGen RC298 Long-Acting TACI-Fc Dual BAFF/APRIL Antagonist Systemic Lupus Erythematosus & Autoimmune Diseases Investigational New Drug (IND) Filing in China & Global Phase 1 Start H1 2027

10. Primary Sources & Verified Citations

Clinical Trials & Regulatory Registries

  • Rezolute Phase 3 upLIFT Study in Tumor Hyperinsulinism: NCT06881992
  • AbelZeta Phase 2 Study of Prizlon-cel in r/r Large B-Cell Lymphoma: NCT05800977
  • BioNTech / Biotheus BNT327 Non-Small Cell Lung Cancer Trial: NCT06841055
  • BioNTech / Biotheus BNT327 Small Cell Lung Cancer Trial: NCT06449209
  • Recursion Phase 2 Sycamore Study of REC-994 in CCM: NCT05085561
  • Recursion Phase 2/3 POPLAR Study of REC-2282 in NF2 Meningioma: NCT05130866
  • Recursion Phase 2 LORIKEET Study of REC-4881 in FAP: NCT05552755
  • Schrödinger Phase 1 Study of SGR-1505 in Mature B-Cell Malignancies: NCT05544019
  • Schrödinger Phase 1 Study of SGR-3515 in Advanced Solid Tumors: NCT06463340
  • Vor Biopharma Phase 1/2 Study of VOR33 in CD33+ AML: NCT04849910
  • U.S. National Library of Medicine Clinical Trials Registry: ClinicalTrials.gov
  • U.S. Food and Drug Administration Regulatory Guidance: FDA.gov

Scientific Preprints & Open-Source Repositories

  • Google DeepMind AlphaFold 3 Official Code & Weights Repository: GitHub AlphaFold 3
  • MIT Jameel Clinic Boltz-1 Biomolecular Modeling Repository: GitHub Boltz
  • Wohlwend, J., Corso, G., Passaro, S., et al. “Boltz-1: Democratizing Biomolecular Interaction Modeling.” bioRxiv: DOI: 10.1101/2024.11.19.624167
  • Retalack, et al. “Natural Language Prompts Guide the Design of Novel Functional Protein Sequences.” bioRxiv: DOI: 10.1101/2024.11.11.622734
  • Rettie, S., Baker, D., et al. “Accurate de novo design of high-affinity protein binding macrocycles using deep learning.” bioRxiv: DOI: 10.1101/2024.11.18.622547
  • Cold Spring Harbor Laboratory bioRxiv Preprint Server: bioRxiv.org

Regulatory Filings (U.S. Securities & Exchange Commission)

Verified Corporate Disclosures & News Feeds