Brandon Richard

$TWST Twist just launched Express Screening Antibodies, a new high-throughput, cell-free service built to help solve a major wet-lab bottleneck for AI drug discovery. AI models can design thousands of antibody sequences in seconds, but testing them in the real world is slow. Previously, researchers had to either order raw DNA (in 4 to 7 days) and run their own cell transfections in-house, or wait 2 to 4 weeks for Twist's living mammalian cell (HEK293/CHO) expression workflows. This new launch builds a fast, automated "middle tier" between raw DNA and full mammalian production. Key highlights: • 5-Day Turnaround: Powered by LenioBio's ALiCE cell-free eukaryotic system, Twist converts digital sequences directly into purified, functional antibodies (5 to 20 µg of full-length IgG, scFv-Fc, or VHH-Fc) starting in just 5 business days. • Built for AI Scale: Engineered to match the massive output of generative AI models, allowing researchers to physically test thousands of antibody candidates per day. • Integrated Binding Data: Purified antibodies are routed straight into Surface Plasmon Resonance (SPR) assays to measure real-world target binding kinetics and affinity right on Day 5. Some context no how this fits into Twist's portfolio: • Synthetic DNA (Clonal/Express Genes): Delivers raw DNA in 4 to 7 days for teams running their own wet-lab expression. • Cell-Free Screening (New Express Screening Abs): Delivers µg-scale purified antibodies + SPR binding data in 5 days to rapidly triage massive AI sequence pools. • Mammalian Expression (HEK293/CHO): Delivers mg-scale purified protein in 10 to 25 days for deep characterization and pre-clinical lead selection. In short, Twist is giving AI biopharma teams a way to filter out non-binders and feed physical binding data back into their AI models in less than a week, before spending the time and money on full mammalian cell expression for their best candidates.

$PRME This guidance seems tailored to Prime's modular platform, even mentions "reverse transcriptase-based editors" twice 👀 Here are 4 key insights and why this could be a massive catalyst for Prime: 1) One BLA Across Mutational Families CBER's new blueprint outlines how companies can bundle multiple mutation-specific product variants for a single gene under one BLA, brand name, and NDC. Under the guidance, sponsors can establish an "Individualized Feature Space" (IFS). The core drug components (like the LNP delivery vehicle, formulation, and editor engine) stay fixed, while the guide sequences vary to address different patient mutations. • Instead of filing separate, costly BLAs for every individual mutation in a disease, the FDA would legally consider these variants to share the same active ingredient. • All variants could share a single brand name, prescribing label, and NDC (National Drug Code), with the container label simply specifying the targeted gene variant. 2) Adding New Mutations Post-Approval via In Vitro Data Alone Once an initial IFS is approved with clinical proof-of-concept, sponsors wouldn't need to run new clinical trials to add unstudied variants. • Under an approved Postapproval Change Management Protocol (PACMP), new mutation-specific guides could be added through streamlined CMC manufacturing supplements using in vitro potency and NGS off-target safety assays. • These in vitro safety packages could even be generated using scaled-down, representative material rather than full commercial batches. 3) Custom CMC and Process Validation (PPQ) Relief Prime editing pegRNAs vary in length and secondary structure depending on the mutation, which traditionally created major manufacturing validation headaches. • CBER explicitly addresses this, instructing sponsors to design flexible chromatography and mixing controls upfront rather than revalidating the process for every sequence. • Even better, the FDA removes the requirement to run Process Performance Qualification (PPQ) campaigns on every variant. Sponsors could bracket manufacturing scales instead. For Prime, managing a $108.8M cash balance as of Q2, avoiding redundant clinical trials and validation campaigns would preserve critical runway. 4) Commercial Blueprint for Wilson Disease and Beyond As mentioned, CBER explicitly cites "reverse transcriptase-based editors" (part of Prime's editing platform) in the text when defining "platform" boundaries. • In Wilson disease, Prime's lead in vivo asset (PM577a) targets the western H1069Q mutation, while a follow-on candidate targets the East Asian R778L mutation. Prime could clinically validate both under a shared ATP7B IFS, secure a single BLA, and then systematically roll out pegRNAs for the long tail of rare ATP7B mutations using in vitro data to capture the entire global market. • The exact same playbook applies to Prime's cystic fibrosis program (covering non-modulator CFTR mutations), PM647 in AATD (expanding from PiZ to rare SERPINA1 variants), and ex vivo CGD. Why is this all important? A major bear thesis on mutation-specific gene editing has always been commercial fragmentation. If a therapy only fixes one specific spelling mistake, how do you affordably treat the hundreds of other mutations in the same gene? This new draft guidance directly answers that question. Keep in mind that this framework applies to multiple mutations within a single target gene, so PM577a (Wilson) and PM647 (AATD) will still require separate BLAs despite sharing the same Universal Liver LNP. In summary, this draft guidance provides the exact regulatory bridge Prime would need to overcome mutation fragmentation and allow them to turn bespoke single-mutation fixes into scalable, single-license commercial franchises without having to fund standalone clinical trials for every variant.

$ABCL Sharing for AbCellera investors following TCEs, pretty cool to see a visual of T-cell Engagers in action 🤓

$TEM Digging into the ESMO 2026 abstract roster shows how deeply embedded Tempus's infrastructure is becoming across clinical oncology research. I found 40 abstracts that mention Tempus. 24 are just standard co-author conflict-of-interest disclosures but 16 are active scientific studies directly using Tempus sequencing panels (xT, xR), liquid biopsy assays (xM/NeXT Personal Dx), AI pathology, and the Tempus Lens multimodal database. Breakdown of 3 key takeaways: 1) Slashing ADC Screening Bottlenecks by roughly 75% with RNA Identifying patients for Antibody-Drug Conjugates (ADCs) targeting Claudin-6 (CLN6) typically requires invasive and costly immunohistochemistry (IHC) protein staining on precious biopsy tissue. UCLA researchers used routine Tempus whole-transcriptome RNA sequencing (xR) across 466 metastatic lung cancer patients and showed that existing RNA data predicts CLN6 protein positivity with near-perfect accuracy (0.976 AUC, 100% sensitivity, 84% specificity). This allows clinics to use Tempus RNA data as an automated pre-screen, cutting the need for physical IHC staining down to only about 25% of tumor samples. 2) Uncovering Novel Biomarkers via Multimodal Data Researchers are actively mining the Tempus Lens database (which pairs DNA "xT" and RNA "xR" sequencing with real-world patient outcomes) to discover what drives treatment resistance and survival: • GCN2 in Prostate & Kidney Cancers: Across 8,456 advanced prostate cancer patients and a companion kidney cancer cohort, high GCN2 RNA expression identified a distinct, less aggressive tumor profile with fewer high-risk mutations (TP53, RB1, AR) and significantly longer time to next treatment (adjusted HR = 0.78, p = 0.044). • MTAP Deletions & Immunotherapy Resistance: Using Tempus's US real-world dataset, an international study showed that MTAP gene deletions occur in roughly 15% of lung adenocarcinomas, remain stable across treatment lines, and serve as a negative biomarker predicting shorter survival on standard checkpoint inhibitors. • Granular Sub-Cohort Insights: Seven additional database studies evaluated early-onset prostate and breast cancers, PTEN and ATM mutations, real-world Enhertu (T-DXd) outcomes in gynecologic cancers, head and neck immune archetypes, and druggable TP53-Y220C mutations across pan-solid tumors. 3) Real-World Liquid Biopsy & Global Expansion • ctDNA vs. Imaging Scans: UCSF and Stanford researchers evaluated Tempus xM/NeXT Personal Dx in 137 recurrent/metastatic head and neck cancer patients, finding a 22% discordance rate between blood ctDNA tests and imaging scans (most often visible disease on scans despite negative ctDNA). This highlights why combining liquid biopsy with imaging and multimodal data is critical rather than relying on a single blood test in isolation. • European Footprint: The multicenter European LOLA Project shared interim real-world results from deploying large-panel ctDNA NGS testing across metastatic lung and colorectal cancer centers in Europe. ESMO 2026 shows that academic centers and biopharma aren't just using Tempus as a commodity lab vendor, they're using xR RNA sequencing to eliminate tissue testing bottlenecks and mining the Lens database to discover the next generation of clinical biomarkers.

$TWST Usual suspect spotted in the RFD4-Proteina scientific preprint, not surprising considering Twist's close relationship with David Baker's lab.

$TEM Two new announcements today from Tempus involving Bristol Myers Squibb (BMS), and Pathos AI that highlight how Tempus is scaling its data moat across both oncology and neuroscience: 1) Expanded BMS Oncology Deal (Through 2031) • Long-Term Data Licensing: BMS expanded its multi-year agreement to license Tempus's de-identified multimodal oncology data through 2031, making BMS one of Tempus’s largest oncology data licensors. • Lens Platform & GPU Access: BMS gains full access to Tempus's Lens Platform, an AI-first, agentic environment connected to dedicated high-performance GPU infrastructure. This allows BMS researchers to directly build, train, and execute advanced AI models on training-ready datasets. • R&D Focus: The collaboration aims to uncover novel drug targets, understand resistance mechanisms, and increase the probability of technical success across BMS's oncology pipeline. 2) Pathos AI & BMS Neuroscience Collaboration ($55M Deal) • Expanding into Neurology: Tempus will use its multimodal infrastructure to build a curated, de-identified longitudinal dataset tracking hundreds of thousands of patients with Alzheimer's and other neurological conditions. Pathos AI is licensing this dataset under a 5-year, $55M agreement. (Important to note that Tempus founder Lefkofsky also serves as the Co-founder and Chairman of the Board of Pathos AI) • BMS Backing & Pathos Foundry: Alongside the data deal, BMS led Pathos's $237M Series E financing at an approximate $3B post-money valuation. BMS secured multi-year access to the Pathos Foundry platform (which uses thousands of specialized AI agents to calculate real-time probability of success, guide trial pivots, and pinpoint patient sub-populations) along with 3-year exclusive rights for its neuroscience programs. Key Takeaways: • Proving the Model Beyond Oncology: A major question has been how effectively Tempus can replicate its oncology playbook in other disease areas. Landing a $55M contract to build out a massive longitudinal neurology dataset is strong validation of its non-oncology expansion. • Dual Monetization Channels: Tempus locks in direct, recurring high-margin licensing revenue from pharma giants like BMS through 2031, while also capturing upside from techbio partners like Pathos through data fees, cloud hosting, and equity stakes. • Deepening Enterprise Integration: By embedding Tempus's Lens infrastructure and Pathos's Foundry platform across its oncology and neuroscience pipelines, BMS is making Tempus's data ecosystem a foundational piece of its daily R&D workflow. To summarize, Tempus is showing that its multimodal data library isn't just a static reference tool for cancer. It's an active, GPU-powered operating system that pharma giants and emerging AI biotechs are paying tens of millions to build on top of, now expanding into neuroscience.

$ABCL AbCellera’s official poster for ABCL635 from TMS 2026 has been released. It consolidates the Ph1 SAD and MAD, and Ph2 proof-of-concept cohorts into a single dataset and gives us 4 new data points: 1) First Look at Immunogenicity (2% ADA Rate) AbCellera disclosed the anti-drug antibody (ADA) rate across the combined Phase 1 SAD (Single Ascending Dose) and MAD (Multiple Ascending Dose) cohorts: just 1 out of 42 participants (2%). More importantly, there was no significant relationship between ADA positivity and drug clearance. This is a key de-risking milestone for a chronic monthly injectable, confirming the immune system isn't clearing the drug early or reducing exposure. 2) Quantified Sleep Improvement (8.9-Point Drop) Earlier updates noted a statistically significant improvement in sleep, and the poster puts hard numbers to the PROMIS sleep disturbance scale. At week 4, ABCL635 delivered an 8.9-point improvement in sleep disturbance compared to a 3.5-point improvement on placebo (a net 5.4-point benefit over placebo). 3) Side-by-Side Repeat Dosing Liver Data (Zero Signals) The poster features side-by-side longitudinal graphs comparing liver enzymes (ALT, AST, bilirubin) from the 3-dose monthly MAD cohort (Sept 16, 2026 cutoff) against the single-dose Ph2 cohort (July 30, 2026 cutoff). Repeat monthly dosing over 3 months maintained the exact same flat, baseline-level liver profile as a single dose, further separating ABCL635 from oral small molecules like Veozah and Lynkuet. 4) MENQOL Quality-of-Life Breakdown On the Menopause-Specific Quality of Life (MENQOL) scale, ABCL635 drove a statistically significant improvement in the core vasomotor symptoms (VMS) domain over time, while other sub-domains (psychosocial, physical, and sexual health) trended favorably vs. placebo without reaching statistical significance. In short, this poster from The Menopause Society 2026 annual meeting checks several of the remaining boxes as ABCL635 heads toward late-stage development. We now have confirmation of a very low (2%) immunogenicity rate that doesn't impact drug clearance, quantified sleep benefits, and flat liver enzyme curves across repeat monthly dosing, backing up its best-in-class potential.

$PRME One of the aspects I like most about Prime's platform is the modularity and how it positions them to capitalize on recent FDA efforts to streamline approvals and make bespoke genetic medicine more scalable. Although prime editing can theoretically repair most genetic mutations, old FDA rules would treat every new guide RNA as a brand-new drug. Spending years and tens of millions of dollars repeating animal tox, biodistribution, and manufacturing validation for single point mutations left thousands of rare diseases commercially stranded. In light of ongoing shifts at the FDA, Prime is well positioned to break through this bottleneck: 1) Turning Drug Development into an Interchangeable Process Prime’s Universal Liver LNP packages five standardized chemical components around prime editor mRNA. • To target a new liver mutation or disease, Prime leaves the entire nanoparticle shell, delivery formulation, and editor mRNA completely intact, swapping only the guide RNA sequences (less than 5% of the drug’s total mass). • Under the FDA’s Platform Technology Designation and Leveraging Prior Knowledge frameworks, Prime could cross-reference safety, stability, and CMC data across programs. This is likely why PM647 (AATD) cleared its IND in September 2026 just weeks after PM577a (Wilson disease) in July, cutting preclinical time and capital expenditure with each successive pipeline asset. 2) "Umbrella" Approvals Across Mutational Families In Wilson disease, Prime’s lead in vivo candidate (PM577a) targets the H1069Q mutation common in the West, while a follow-on candidate addresses the R778L variant prevalent in East Asia. • Under the FDA’s draft Plausible Mechanism Framework, once PM577a demonstrates clinical proof-of-mechanism in the human liver, Prime could potentially bundle follow-on ATP7B guides into master protocols or supplemental filings without running multi-year, standalone trials for every variant. • As CEO Dr. Allan Reine noted after dosing the first PM577a patient this month, the same shared LNP delivery approach provides a direct path to expand into global patient populations across mutational families. 3) De-Risked Human Biology & Near-Term Catalysts Prime editing is no longer purely theoretical. In PM359 for p47phox CGD, Prime proved in humans (NEJM, Dec 2025) that prime editors repair DNA cleanly without double-stranded breaks, restoring functional neutrophil activity to 69–83% (well clear of the 20% clinical efficacy threshold). • Holding RMAT and Rare Pediatric Disease designations, PM359 is on track for a streamlined 1H 2027 BLA under the FDA's Rare Disease Evidence Principles (RDEP), a milestone that could unlock a Priority Review Voucher (PRV) worth about $100M+ in non-dilutive capital. • Upcoming 2027 clinical readouts for PM577a (Wilson) and PM647 (AATD) rely on direct, objective biomarkers (serum ceruloplasmin, 64Cu PET copper efflux, and functional M-AAT protein), insulating the platform from the subjective clinical-scale failures that tripped up others. In short, the modular nature of Prime's platform combined with the FDA's regulatory updates could allow the tech to evolve from an unscalable trial-and-error grind into a plug-and-play platform, sharply lowering the cost and latency required to commercialize genetic cures.

Eli Lilly CEO Dave Ricks says AI helping us conquer disease will require "lots and lots of data that doesn't exist, yet." Why I continue to focus on companies that own properietary data (like AbCellera & Tempus) and companies that help generate new data at scale (like Twist).

Same. Although, I wouldn't mind if Twist and Adaptyv combined forces 😏

More AI x Bio 👀

$ABCL AbCellera’s official abstract for ABCL635 from The Menopause Society 2026 meeting is out. While we already knew the high-level Phase 2 data, this new TMS abstract gives us a deeper look at a few more granular data points: 1) Exact Headache Breakdown (Clustered at the Highest Dose) Prior slides noted mild headache as a common adverse event, but we now have the exact numbers. In the SAD cohort, transient, mild headache occurred in 23.3% (n=7) of participants within the first few days of dosing. Importantly, more than half of those cases (n=4) were clustered at the highest 900mg supratherapeutic dose, well above the 300mg and 600mg doses taken into the multi-dose cohort. 2) Quantified Target Engagement (65% to 80% Testosterone Drop) In Q1 earnings slides, AbCellera showed line graphs of hormone suppression to prove ABCL635 was hitting the target. The abstract puts more hard numbers to those charts: male participants saw a 65% to 80% reduction in testosterone from baseline across the 300mg to 900mg doses (alongside corresponding drops in LH and FSH), confirming strong, concentration-dependent HPG axis suppression. 3) Cohort Demographics The abstract also discloses the exact median ages for the trial groups: 54.5 years for the SAD cohort and 62.0 years for the MAD cohort. Quick recap of the Phase 1/2 core results: • Trial Setup: The Single Ascending Dose (SAD) cohort enrolled 40 healthy men and postmenopausal women (6:2 active-to-placebo across 30mg, 100mg, 300mg, 600mg, and 900mg). The Multiple Ascending Dose (MAD) cohort enrolled 16 healthy postmenopausal women (6:2 active-to-placebo receiving 3 monthly doses of 300mg or 600mg), alongside the Phase 2 proof-of-concept cohort in women with moderate-to-severe VMS. • Clean Safety & Liver Profile: Zero serious or severe adverse events, no dose-limiting toxicities, and (crucially for the NK3R class) no clinically significant liver enzyme elevations across either the SAD or MAD cohorts. • Validated Monthly Dosing (PK): Dose-proportional exposure with a roughly 24-day half-life, backing up the convenient once-every-4-weeks injection schedule. To summarize, the granular numbers reinforce the bull case for ABCL635. We’re seeing potent, quantifiable target engagement at therapeutic doses, a roughly 24-day half-life that easily supports monthly dosing, and a very clean safety profile where even mild side effects like headaches were largely confined to the maxed-out 900mg test dose, all with zero liver toxicity signals.

$TWST More collaboration between Twist and Anthropic: "Teams at Adaptyv Bio and Twist Bioscience experimentally tested 211 designs for binding." Plus, Citrini launches Citrini Therapeutics 👀

$TEM Tempus presented new clinical validation data at ESMO 2026 for an AI-enabled pre-screening panel in advanced non-small cell lung cancer (NSCLC), developed through its ongoing collaboration with Pfizer. As the number of targeted therapies for lung cancer grows, identifying the right biomarkers requires running multiple, time-consuming tests on limited biopsy tissue. In practice, tumor tissue often runs out before testing is finished, forcing patients to go through invasive repeat biopsies or delaying targeted treatment. Tempus is addressing this diagnostic bottleneck by pulling molecular signals directly from standard hematoxylin and eosin (H&E) digital pathology slides that are already generated during routine care. Key takeaways: • Massive Real-World Validation: The multi-center analysis evaluated 60,459 advanced NSCLC cases across 56,969 patients using routine whole-slide pathology images. • High Analytical Sensitivity: The AI panel achieved more than 80% sensitivity across all 12 evaluated genomic and IHC biomarkers, with sensitivity exceeding 90% for several specific targets. • Intelligent Diagnostic Triage: Still in development, the panel isn't meant to replace established sequencing (NGS) or immunohistochemistry (IHC) assays. Instead, it acts as a triage tool to help clinicians prioritize which confirmatory tests to order first, preserving limited tissue samples and getting patients on targeted therapies faster. • Pharma Alignment: Conducting the study with Pfizer highlights a clear shared incentive. Pharma companies want to remove testing hurdles so more eligible patients actually get identified for precision therapies. • Expanding the Digital Pathology Suite: This builds directly on Tempus’s existing AI biomarker work, sitting alongside Paige Predict (which analyzes H&E slides to predict 123 biomarkers across multiple cancer types when tissue is insufficient for sequencing) and Tempus Preview (which provides early biomarker predictions about 24 hours after a sample arrives at the lab). In short, Tempus is using its multimodal data library to turn a routine pathology slide into an automated triage tool, helping doctors conserve tissue, eliminate testing delays, and route patients toward personalized treatments much faster.

$PRME Regulation can make or break a thesis in bio. Between 2024-2026, the FDA has been overhauling how it evaluates genetic medicines and platform technologies like Prime editing, creating a potential tailwind for this type of tech. Roughly 95% of rare diseases still have no approved treatment. Instead of forcing ultra-rare programs into slow, legacy clinical trial models designed for mass-market pills, the FDA has announced several policy changes to modernize the process: • Plug-and-Play Platform Reuse (May 2024 & Summer 2026) Once a delivery vehicle or manufacturing method (like a standardized LNP or AAV vector) is approved, sponsors no longer have to start from scratch. Under Section 506K and recent genome-editing guidance, companies can cross-reference prior safety, biodistribution, and manufacturing (CMC) data across different drug candidates using the same platform. • Single-Arm Approvals for Ultra-Rare Diseases (Sept 2025) For severe genetic conditions affecting less than 1,000 U.S. patients, randomized placebo trials are practically impossible. Building on previous policy, the FDA’s Rare Disease Evidence Principles (RDEP) now permit full marketing approval based on a single-arm trial, provided it’s backed by clear natural history data and confirmatory biomarker readouts. • Mutation Bundling via Draft "Plausible Mechanism" (Feb 2026) Gene editing therapies often target distinct, patient-specific mutations within the exact same gene. Under the Plausible Mechanism Framework, sponsors could bundle multiple mutation-specific editors into a single trial and BLA under a master protocol. In addition, new mutation variants could potentially be added post-approval without requiring full-scale clinical trials. • Restored Non-Dilutive Capital & Division-Wide Alignment (2026) Congress and the White House reauthorized the Rare Pediatric Disease Priority Review Voucher (PRV) program, restoring an asset historically worth about $100M+ in cash upon approval. Capping this off, the FDA’s recent rare disease framework whitepaper standardizes Bayesian stats and flexible CMC inspections across all CDER and CBER review divisions to eliminate inconsistent standards. In short, for modular genetic medicines with hard biomarker readouts, the FDA is starting to dismantle the traditional 10-year development gauntlet in favor of a faster, capital-efficient, plug-and-play framework. While some of these shifts are still in a draft or non-binding guidance status, they're shaping up to be a regulatory sweet spot for therapies like prime editing that directly repair a root genetic mutation and deliver objective, binary biochemical proof.