Announced June 8 at Wellcome Sanger’s AI x BIO conference, the consortium aims to generate high-quality, AI-ready genomic datasets designed to train advanced machine learning models. The datasets will provide a foundation for training and evaluating AI models capable of making more accurate predictions about biological processes, the partners said.

Google DeepMind and Google.org, Google’s philanthropic arm, have committed $5 million per year over five years.

The initiative builds on an existing relationship between the two organizations, which includes previous research collaborations and a jointly-sponsored AI in genomics fellowship.

“Together with the Sanger Institute, we aim to build the data backbone needed to decode the complexities of biological processes,” Pushmeet Kohli, VP of AI for Science at Google DeepMind, said in a statement. “Ultimately, this could accelerate scientific discovery and unlock entirely new frontiers for researchers worldwide.”

Julia Wilson, chief innovation and impact officer at Wellcome Sanger, added that they plan to share their data resources with the wider scientific community.

Wellcome Sanger and Google DeepMind plan to bring in additional collaborators, they said.

Continuation of the partnership beyond initial milestones is subject to terms set out in the agreement.

Under the agreement, announced June 1, the partners will work in four “concrete” areas: jointly advancing clinical implementation of WGS as a default diagnostic tool for adults with acute leukaemia; validating and implementing Hartwig’s OncoAnalyser cancer genomics analysis pipeline for hematological malignancies; benchmarking and progressively rolling out WGS for selected solid tumours; and harmonizing the data models used by both organisations for genomic and clinical data.

Financial terms were not disclosed.

“This partnership is another important step towards making precision oncology a standard, equitable part of cancer care,” Richard Brandell, Director of Genomic Medicine Sweden, said in a statement. “By aligning our efforts with Hartwig, we accelerate the journey from genomic insight to better outcomes for patients in Sweden and beyond.”

The deal combines Sweden’s nationwide infrastructure for genomic diagnostics with Hartwig’s sequencing capabilities, bioinformatics pipelines, and database.

The collaboration will also consider data already generated through routine care. The organizations aim to make their datasets comparable and reusable for research while enabling the development of new diagnostics and therapies. Longer term, they intend to support federated analyses, enabling researchers and clinicians to query Swedish, Dutch, and future partner datasets without sensitive data leaving its national jurisdiction.

The proposed hub, announced June 4, would combine MSKCC’s clinical data and scientific expertise with Sophia’s SOPHiA DDM platform, which the company describes as an AI and analytics layer for precision medicine. Under the envisioned structure, SOPHiA DDM would serve as the hub’s AI, data, and bioinformatics backbone, while MSKCC would contribute its oncology faculty, clinical leadership, and capacity to generate high-quality patient data.

Specific terms of the joint venture have not yet been finalized, the partners said in a statement, and financial details were not disclosed.

“This collaboration represents the future of applied precision oncology – where clinical expertise meets cutting-edge Ai and robotics with the potential to transform cancer care at scale,” Michael Frank, Director of Digital Health Business Development at MSKCC, said in a statement:

Both institutions bring substantial genomic datasets to the arrangement. SOPHiA DDM has been used to analyze more than 2.5 million cases since 2014, while MSK’s in-house next-generation sequencing program has sequenced more than 150,000 tumor samples over the same time frame, building a dataset linked to clinical outcomes, pathology, and radiology records.

The deal builds on an existing relationship between the entities. Through prior licensing agreements covering sequencing-based cancer panels including MSK-IMPACT and MSK-ACCESS, Sophia has distributed MSK diagnostic tools in 35 countries.

MSKCC and Sophia suggested that their biopharma partners would gain access to infrastructure for companion diagnostic development, clinical algorithm development, and evidence generation.

No timeline for finalizing the joint venture agreement was disclosed.

That’s how my guest this week, Arc Institute Senior Scientist Alexander Nevue, describes PerturbSpace, a new method he and his colleagues introduced in a May 26 BioRxiv preprint.

PerturbSpace is the latest method to marry the power of CRISPR-based perturbation screening, where genes across the genome are systematically targeted and the effects of those perturbations are read out in the transcriptome of the cell in which they were made.

This juncture of CRISPR-based screening and spatial transcriptomics is one of the hottest areas in molecular biology right now, with nearly 10 different methods published in the last two years, from leading scientists including Genentech’s Aviv Regev; Harvard University’s Xiaowei Zhuang, the inventor of MERFISH and cofounder of Vizen; Yale University’s Rong Fan, and the Broad Institute’s Sami Farhi.

On a personal note, this discussion brought me back to when I was covering the flurry of CRISPR-based technologies that were coming out a decade ago. I reported on the first perturb-seq papers when they came out in 2016 (!) and its fascinating to see it intersect with two more recent trends I’ve been following: spatial biology and AI virtual cell development.

Join Alex and I as we discuss the PertubSpace workflow, its advantages and disadvantages compared to other spatial perturb-seq methods, and where he sees the data from these screens making the most impact.

You can also listen to the interview on Apple Podcasts and Spotify

Subscribe now

Links to other publications of spatial perturbation screening methods with notable authors:

PerturbView. Aviv Regev and Erik Lubeck of Genentech. https://www.nature.com/articles/s41587-024-02391-0

Large-scale pooled in vivo CRISPR screening. Jonathan Weissman of MIT and Harvard’s Xiaowei Zhuang, inventor of MERFISH and cofounder of Vizgen. https://www.biorxiv.org/content/10.1101/2024.11.18.624217v3

SPAtial Cell Exploration (SPACE). Joe Beechem of Bruker/NanoString, in collaboration with Merck. https://www.biorxiv.org/content/10.1101/2025.09.14.675819v2

Perturb-DBiT. Rong Fan of Yale University, inventor of DBIT family of methods and cofounder of AtlasXOmics. https://www.researchsquare.com/article/rs-6481967/v1

CRISPRMap. Jellert Gaublomme of Columbia University, Dan Landau of Weill Cornell Medicine. https://www.nature.com/articles/s41587-024-02386-x

Perturb-FISH. Sami Farhi of the Broad Institute. https://www.cell.com/cell/fulltext/S0092-8674(25)00197-7

NIS-Seq Jonathan L. Schmid-Burgk of University Hospital Bonn. https://www.nature.com/articles/s41587-024-02516-5

Spatial Perturb-Seq. Wei Leong Chew of A*STAR GIS. https://www.nature.com/articles/s41467-026-69677-6

Under the agreement, announced June 4, Natera will apply its Signatera assay — a tumor-informed, sequencing-based test for minimal residual disease (MRD) — to clinical samples from CytoDyn’s CLOVER study, which is evaluating leronlimab in combination with trifluridine/tipiracil and bevacizumab in previously treated mCRC patients.

“Signatera has become an increasingly important tool in precision oncology and clinical development,” CytoDyn CEO Jacob Lalezari said in a statement. “Through this collaboration, we expect to gain valuable insights into ctDNA response kinetics and disease progression that may help guide future development strategies for leronlimab in colorectal cancer and potentially other solid tumor indications.”

Financial details of the deal were not disclosed.

Natera will also conduct real-world data analyses using its oncology database, which the company says contains more than 2 million plasma timepoints alongside clinical and imaging records. By combining MRD testing data with electronic medical record information, Natera aims to generate analyses of patient populations, treatment patterns, ctDNA response rates, and response dynamics across clinical settings.

The companies said the collaboration is intended to generate molecular insights that could inform future development of leronlimab, including clinical trial design, biomarker-driven patient selection, and translational research.

The round included participation from Almi Invest, Life Science Invest, and SU Ventures, joined by an undisclosed list of other investors.

The company said it will use the funding to begin commercializing its spatial biology technologies, which are designed for clinical validation and translational workflows. The financing also supports collaborations with academic and industry partners.

“Discovery is no longer the bottleneck,” Haga Bio Cofounder Mats Nilsson said in a statement. “The next major step for the field is translating and validating these discoveries at scale and ultimately into clinical use. That is precisely what Haga Bio aims to enable.”

The company’s platform enables sensitive and specific detection of single nucleotide variants (SNVs) in RNA in situ. The methods were developed by cofounders Hower Lee, who serves as CEO; Marco Grillo, who serves as CSO; and Nilsson, a professor at Stockholm University.

Founded in 2024, Haga Bio has designed technologies to validate RNA biomarkers in large clinical studies while remaining cost-effective and scalable. The company said it has now raised a total of approximately $3 million.

Additional co-founders include Malte Kuhnemund, Daniel Gyllborg, and Xiaoyan Qian, all veterans of 10x Genomics. Kuhnemund and Qian cofounded Cartana in 2017, a SciLifeLab spinout whose in situ RNA sequencing technology was acquired by 10x Genomics in 2020 and integrated into its Xenium spatial transcriptomics platform.

Haga Bio announced an early access program in May. Through the program, the company is providing access to two products: Haga Pattern, a multiplexed spatial gene expression assay, and Haga Point, a spatial assay for in situ SNV detection on RNA.

“Until now, the field has lacked robust in situ variant calling on FFPE tissue samples with high sensitivity and specificity,” Lee said in a statement. “Addressing this challenge enables new possibilities for translational research.”

Not content to simply supply sequencing for a sequencing-intensive cancer test, Illumina announced May 28 that it will be launching a new kit for minimal/molecular residual disease. It’s a research-use test for now, however, with the ability to do both solid and blood cancers, it could offer large studies an alternative to sending out to one of the many MRD testing providers. As I mentioned in a newsletter a few weeks ago, it seems as though solid tumor MRD testing is about to explode.

“Illumina’s MRD solution for clinical research leverages the advanced sensitivity of whole-genome sequencing (WGS,) coupled with unparalleled analysis, to enable our customers to more easily deliver the most precise information to advance MRD research,” Todd Christian, senior VP of Services, Arrays, and Genomic Access at Illumina, said in a statement. “We aim to make WGS in oncology more accessible and scalable to support the integration of precision solutions into the standard of care.”

By now, Illumina can be counted on to try and get in on any lucrative application developed using its sequencing tech. Non-invasive prenatal testing (NIPT) may have been the first, when the company acquired Verinata Health in 2013. More recently, Illumina has elbowed into comprehensive genomic tumor profiling with its TruSight Oncology panels; multi-cancer early detection testing, with its (failed) Grail acquisition; single-cell sequencing with its 2024 Fluent BioSciences deal; and soon, spatial transcriptomics.

The kit is already in early access and the end-to-end research workflow can be completed in as fast as 5 days and is optimized for analytical sensitivity as low as 10 ppm.

That wasn’t all Illumina was up to this week. It also launched automated library prep workflows for the MiSeq i100 benchtop instrument on the Firefly Go with SPT Labtech.

“This solution also enables automation capabilities for a growing menu of NGS panels developed by Pillar Biosciences, supporting Illumina’s targeted oncology research portfolio,” Illumina said in a statement. Important, because it is also expanding its menu of Pillar’s oncoReveal panels to include the oncoReveal Nexus 21 Gene Panel for rapid solid tumor and hematological malignancy sequencing and the oncoReveal Lymphoid Panel, an 84-gene DNA panel.

Moreover, Illumina partnered with Danaher’s IDT to create DRAGEN data analysis for certain IDT hybridization-based library prep workflows. Initial workflows focusing on FFPE samples are available now. A version optimized for cfDNA liquid biopsy targeted applications is slated for next year.

The deal focused on research workflows, however, this year IDT announced it is heading into the clinical assay market, so having data analysis pathways already established in Illumina’s ecosystem could help it attract less bioinformatically sophisticated labs.

Subscribe now

Non-Illumina News

Researchers from the University of Cambridge have published a study in Science describing how they mapped cells that produce and receive hormone signals.

“Drawing inspiration from the Human Cell Atlas, we analyzed expression of 379 hormone and receptor genes in a transcriptomic dataset comprising 14 million single cells and nuclei across 47 human tissues,” the authors wrote. “We predicted non-classical sites of hormone expression, including secretin in plasmacytoid dendritic cells, inferred convergent hormone action and endocrine feedback loops, and implicated cell populations in monogenic endocrine disorders.”

Data is available to view and download at https://hormonecellatlas.cellgeni.sanger.ac.uk/

ASCO Not Impressed by Grail Galleri Trial Data, Won’t Recommend Test

The American Society of Clinical Oncology (ASCO) won’t be recommending the Grail Galleri multi-cancer early detection test, a spokesperson said at the society’s annual meeting, which wraps today.

As reported by Kelsy Ketchum and Turna Ray at GenomeWeb, Julie Gralow, the organization’s chief medical officer and executive VP, said “at this time, this Galleri assay should not be included in cancer screening guidelines.”

In February, I reported on Grail’s disclosure that its 142,000-participant UK NHS-Galleri trial had not met its primary endpoint of a reduction in Stage III and IV cancers. At ASCO, Grail presented complete data from that trial as well as its PATHFINDER 2 study.

Canaccord Genuity Analyst Kyle Mikson wrote in a May 31 note to investors that the “FDA’s review [of Galleri] will be driven by the promising safety and clinical performance data from PATHFINDER 2 and NHS-Galleri (rather than clinical utility data).” Moreover, he said he believes Medicare coverage decisions will emphasize US-based studies.

What I’m Reading

Murphy, M. and O’Neill, L. “A break in mitochondrial endosymbiosis as a basis for inflammatory diseases.” Nature (2024).

Why do signaling pathways triggered by infection or tissue injury use mitochondria as a central hub?

“Whereas mitochondria can be considered ‘the enemy within’ the cell, evolution has used this strained relationship in intriguing ways, with increasing evidence pointing to the recent failure of endosymbiosis being critical for the pathogenesis of inflammatory diseases.”

Elsewhere on the Internet

Hundreds of papers — some published in journals such as Nature and Immunity, by researchers at institutions including the Salk Institute and University of Cambridge — have potentially incorrect data because the Abcam website brings up an antibody for the wrong protein when you search for the p16 tumor suppressor.

Last week I wrote about data sleuth Sholto David’s discovery that the Thermo Fisher Scientific website had allegedly manipulated Western blots in its validation data for certain antibodies. Now, he writes that this p16 mix up has seeped into the literature.

“There is something unsettling about the idea that hundreds of papers can be published using completely the wrong antibody without anyone noticing, as if the research results are only loosely connected with biological reality. How many antibodies could you swap for loading controls until someone actually takes notice of what the results are telling them?” David wrote.

It’s a Toy Story, baby just say ‘Yes’

Taylor Swift announced this week that she has recorded a new, original song for the Toy Story 5 soundtrack, titled “I Knew It, I Knew You.”

“I’ve always dreamed of getting to write for these characters who I’ve adored since I was a 5 year old kid watching the first Toy Story movie,” Swift wrote in a June 1 Instagram post.

Kudos to the Disney execs who found a way to get more millennial parents into movie theaters when the movie comes out June 19. If you don’t want to see Buzz Lightyear battle an iPad, you can just listen to the song this Friday, June 5.

Illumina built a new application for IDT custom panels and includes a low-quality FFPE mode designed to extract variant data from degraded samples. Initial workflows focusing on FFPE samples are available now. A version optimized for cfDNA liquid biopsy targeted applications is slated for next year. Financial terms of the collaboration were not disclosed.

“This is a step toward reshaping how genomic data is generated, analyzed, and used in research,” Konstantin Fiedler, VP at IDT, said in a statement. “We’re building toward a future where more workflows are fully connected, highly performant, and built for real-world samples.”

The deal adds to already deep ties between the companies: IDT manufactures oligos used in Illumina’s sequencing-by-synthesis chemistry. While the current deal is centered on research-use library preparation workflows, earlier this year IDT launched its first in vitro diagnostic assays, the Archer FusionPlex-HT Dx and VariantPlex-HT Dx, both of which can use FFPE samples.

“Collaborations like this add a new way for people to enter Illumina’s insight ecosystem,” Kevin Moore, VP of software product management at Illumina, added.

Over the last two weeks, scientific misconduct sleuths have uncovered what appears to be systematic image manipulation in antibody validation data posted to the website of Thermo Fisher Scientific, the world’s largest supplier of instruments and reagents.

Sholto David, a UK-based researcher, found the first instance of doctored Western blots and others found many more.

“The images are 100 percent improperly edited,” David told Ion Genomics. “Some of the most obvious I’ve seen.”

That doesn’t mean the antibodies don’t work, he noted, but these images are often all a researcher has to go on when choosing which reagents to order. Thermo Fisher did not respond to multiple requests for comment.

Working on this story brought me back to my stint at Retraction Watch, where I reported on many retractions for image manipulation, as well as False Claims Act lawsuits.

Usually the stakes are lower — the professional currency of scientific publication — though millions of dollars in research grants can also be at play. Recently, Dana-Farber Cancer Institute settled a whistleblower lawsuit brought by David for $15 million, netting him a $2.6 million bounty. To see evidence of image manipulation at a company that did nearly $45 billion in revenue last year was a shock.

David told me he isn’t currently planning on launching a qui tam suit around this. Still, I will have to go back through my emails and try to find a legal expert to see if Thermo Fisher is at risk if somebody somewhere bought these antibodies with federal grant money.

Leave a comment

The week in genomics

Insilico Medicine and a new spinout from Human Longevity will codevelop AI foundation models aimed at decoding the biology of aging.

Insilico Medicine will contribute model architecture, benchmarking, training guidelines, and computational algorithms. Human Life Foundation Models will integrate those tools with Human Longevity’s de-identified multiomic and clinical datasets.

The companies said the jointly developed models are intended to support early detection of age-related diseases, predictive health risk modeling, and the discovery of therapeutics and personalized interventions.

Parse Biosciences and Bit.bio teamed up to map transcription factors contributing to cell identity, with the goal of advancing AI models for use in drug discovery and human cell manufacturing. No financial details of the partnership were disclosed.

Guardant Shares Spiked on FDA approval of Guardant Liquid CDx

Guardant Health received FDA approval for Guardant360 Liquid CDx, a blood-based comprehensive genomic profiling test that assesses a wider genomic footprint than the company’s previously approved Guardant360 CDx.

The new test integrates genomic and epigenomic profiling from a single blood draw, providing increased sensitivity for circulating tumor DNA (ctDNA) detection compared with the prior test. The seven companion diagnostic indications previously approved for Guardant360 CDx transfer to the new test under the approval.

PacBio Begins Shipping Reusable Sequencing Chips

PacBio said it has begun shipping multi-use sequencing reagents for the Revio instrument, based on the Sprq-Nx chemistry. Using these reagents, customers running more than 5,000 genomes per year can achieve a price per human genome (at 20X coverage) of less than $300, according to PacBio.

“In our beta testing, we saw consistently strong run performance,” Adam Ameur, a bioinformatician at Sweden’s Uppsala University, said in a statement. “The simple workflow, low failure rates, and substantially lower pricing with multi-use SMRT Cells make SPRQ-Nx a practical upgrade for large-scale sequencing projects.”

PacBio noted that it has also made improvements to DeepConsensus, a basecalling algorithm developed in collaboration with Google.

“The quality of genomic data is determined by the information richness of the sequencer and the refinement of the algorithms that process it,” Andrew Carroll, Product Lead for Genomics at Google Research, said in a statement. “New advances in DeepConsensus unlock even more of the exceptional quality inherent in HiFi sequencing, empowering scientists and clinicians to find new insights and resolve complex cases of rare disease.”

Natera to Build Huge Sequencing Facility

Natera said on May 21 it would build a new, dedicated sequencing facility in Texas that would give it the “largest sequencing capacity in the world.”

The expansion is driven by demand for its tests, particularly in oncology where it saw an increase in testing volume of 54 percent year over year in the first quarter of 2026.

Natera did not respond to questions about how many additional sequencing instruments it would house in the new facility.

Illumina Adds David King to Board

Illumina elected its latest slate of directors, adding David King, former CEO of Labcorp from 2007 to 2019. King “marks an important addition to Illumina’s Board as the company accelerates its clinical strategy to integrate genomics more broadly into the standard of care,” Illumina said in a statement.

Elsewhere on the Internet

Liverpool FC bid farewell to forward Mohamed Salah and left back Andy Robertson over the weekend at the close of the English Premier League season.

Seemingly able to score at will, Salah dazzled on the right wing during Liverpool’s recent golden age. I will choose to remember him as part of the three-headed-monster of Salah, Sadio Mane, and Roberto Firmino. Salah leaves with the third-most goals scored in club history and the fourth-most goals in the history of the Premier League.

Doing his damage as an attacking defender on the other side of the field, Robertson was arguably the heart and soul of the team during his prime — a tenacious player who wasn’t afraid to wind up anybody, even Lionel Messi.

On May 17, Sholto David, a UK-based researcher and scientific misconduct sleuth, posted on social media that Thermo Fisher’s website “appears to show a fake Western blot for the validation of one of their p53 antibodies [...] This does not appear to be one of the ‘published figures,’ but their own internal data.”

His post included an image suggesting that one of the bands in the Western blot had been flipped or rotated around various axes and pasted into multiple other lanes.

As of May 27, many of the allegedly manipulated images remained on the Thermo Fisher website. The company did not respond to multiple requests for comment.

David’s post sparked others to search Thermo Fisher’s website, and they found similar patterns across multiple product lines. In total, multiple antibodies — including four p53 variants as well as Tau, Retinoblastoma protein, NLGN1, and Caspase 7 antibodies — showed evidence of image manipulation.

What implications the allegedly faked validation data could have isn’t clear. “The images are supposed to show that the antibodies bind specifically to their targets,” David said, noting that it’s possible the reagents do actually work.

“If people manipulate images they usually do that to make the data show something it doesn’t. In this case, I suppose the blots do not come out so clean? Perhaps there are some minor bands or unexpected binding in negative controls? From commentary online it seems like people have good experiences with the antibody clone that I first mentioned, I doubt it’s just totally false,” he said.

Thermo Fisher may not be alone in providing questionable data for its antibodies. “I’m working on something else [...] that absolutely will show widespread confusion/fraud related to antibodies sold by Abcam, stay tuned,” David wrote in response to a comment on LinkedIn.

Image manipulation of Western blots and other scientific data have plagued the peer-reviewed literature and federal grant applications for years. The consequences can be tangible: last year, Dana-Farber Cancer Institute settled a False Claims Act lawsuit for $15 million after David discovered image manipulation in papers and grant proposals submitted to the National Institutes of Health by researchers at the institute. As a whistleblower, David secured a $2.6 million payout.

Whether Thermo Fisher could be subject to a similar suit — say, if federal grant funding had been used to purchase these antibodies — is also unclear. For now, David said he hasn’t considered filing such a suit.

What’s certain is that his most recent sleuthing has struck a nerve, with dozens of comments and over a thousand reactions on LinkedIn alone.

“This explains the need to try 3+ antibodies and do validation by KO to get reliable data,” Kirill Bersuker, a scientist at the Google-associated company Calico Life Sciences, wrote in reply to one of David’s LinkedIn posts. “That’s thousands of dollars and days of research time.”

David speculated that without other quality indicators available “you might just choose the one with the prettiest blot on the supplier’s website. For that reason I can see why people feel pissed.”

The companies said on May 27 that they will combine Parse’s Evercode single-cell sequencing technology with Bit.bio’s opti-ox cell programming platform and its Cell Foundry. Together, they plan to test thousands of genetic variables simultaneously to establish causal links between specific genetic changes and biological outcomes.

“Cells operate on code, and by mapping how specific transcription factors dictate cell fate, we are unlocking that operating system. This collaboration doesn’t just generate data; it provides a foundational map for bit.bio to scale human-relevant models and feed predictive AI systems,” Przemek Obloj, CEO of Bit.bio, said in a statement.

Financial and other terms of the deal were not disclosed.

Bit.bio, which spun out of the University of Cambridge in 2016, has raised more than $200 million to date.

Parse, a University of Washington spinout, was recently acquired by Qiagen for $225 million in cash with additional potential milestone payments.

Insilico Medicine will contribute model architecture, benchmarking, training guidelines, and computational algorithms. HLFM will integrate those tools with Human Longevity’s de-identified multiomic and clinical datasets.

The companies said the jointly developed models are intended to support early detection of age-related diseases, predictive health risk modeling, and the discovery of therapeutics and personalized interventions.

In a statement, the partners described the deal as a “multi-million-dollar collaboration,” however, they did not provide more detail and no timeline for commercial availability of the models was disclosed.

“Human longevity and healthspan represent one of the most complex challenges in biology, “ Alex Zhavoronkov, CEO and founder of Insilico Medicine said in a statement. “By combining Insilico’s expertise in generative AI drug discovery and multimodal foundation models with HLFM’s unique datasets [...] we aim to build a next-generation AI system capable of decoding the biology of aging.”

Insilico Medicine said it has published more than 50 research papers on aging, longevity biomarkers, and related therapeutics since 2014. Insilico also said it recently formed a longevity advisory board chaired by Andrew Adams, group VP of molecular discovery at Eli Lilly.

Earlier this year, Insilico Medicine and Eli Lilly signed a development agreement with a $115 million up-front payment and could be worth up to $2.75 billion.

The Palo Alto, California-based company said on May 20 that the new test integrates genomic and epigenomic profiling from a single blood draw, providing increased sensitivity for circulating tumor DNA (ctDNA) detection compared with the prior test. The seven companion diagnostic indications previously approved for Guardant360 CDx transfer to the new test under the approval.

The test is “the largest FDA-approved liquid biopsy panel and covers a 100-fold larger genomic footprint versus the prior tissue-based G360 CDx,” Canaccord Genuity Analyst KYle Mikson wrote in a May 20 note to investors. “The approval also expands future companion diagnostic opportunities, which could drive higher-margin biopharma revenue and support broader adoption of [Guardant’s Reveal circulating tumor DNA monitoring tests] as a complementary monitoring assay. More broadly, the approval advances Guardant’s Smart Platform and multiomic strategy, while potentially supporting […] portfolio simplification.”

Since the close of the market on May 19, shares of Guardant are up 20 percent at $118.01.

Guardant360 Liquid CDx is intended to support treatment selection decisions for patients with advanced cancer. The test is already approved as a companion diagnostic for multiple therapies in non-small cell lung cancer and colorectal cancer, and is the only FDA-approved companion diagnostic for targeted therapy in advanced breast cancer patients with ESR1 mutations, according to Guardant.

Guardant noted that its Smart Platform, a multiomic technology platform that the company said integrates genomic and epigenomic data, now underlies its full product portfolio.

“Guardant can now apply for [Medicare’s] Advanced Diagnostic Laboratory Test status for Guardant360 Liquid CDx, which comes with favorable pricing opportunities,” Guggenheim Securities Analyst Subbu Nambi added in a note to investors. “Based on management’s commentary, this is what we would expect them to pursue next.” Guardant could also sell the blood-based and tissue-based tests together as well as streamline ordering for blood-based tests, she said.

The FDA’s Signatera decision marks the first time a blood-based molecular MRD test has been cleared as a companion diagnostic (CDx.) It allows clinicians to use the test — which essentially says whether the tumor has returned following initial treatment — to guide treatment decisions such as initiating immunotherapy in positive patients and potentially deferring treatment in those who test negative.

Approval as a CDx means the test is now a prerequisite to prescribing this particular drug, but in the bigger picture for Natera, this week marks the moment that Signatera transitions from a promising lab-developed test with solid evidence behind it into a mandated, reimbursement-protected component of a standard treatment pathway. Now, the company can try to replicate this success across the other cancer types that it has designs on penetrating, including breast and ovarian, lung, and colorectal cancers.

This is a big deal for MRD testing in solid tumors, which is driven by detecting tumor DNA fragments in blood. This class of test is poised to become integral to treatment of solid tumors, which account for the majority of new cancer cases in the US.

When I was at the American Association for Cancer Research meeting last month, I attended a keynote session on the use of MRD in solid tumors that opened my eyes to the utility of these tests. I had previously considered them as a nice surveillance tool to catch recurrence. This is partly true: if your blood-based molecular signal returns, you’re extremely likely to have cancer recurrence. (This is helped by the fact that tests like Signatera are “tumor informed,” meaning they sequence a patient’s particular tumor with its particular mutations, a kind of a molecular “most wanted” poster.)

What I didn’t understand is that for many solid tumors, surgery alone is enough to drive remission in something like 50 percent of cases. Still, those 50 people get treated with toxic chemotherapy because studies show that overall survival is improved when 100 out of 100 people get that chemo. If you get an MRD test and it shows no signs of cancer, maybe your doctor is more comfortable in not prescribing that chemo.

If your tumor is going to come back, however, you may want a more aggressive treatment option provided by immunotherapy. For MIBC, that immunotherapy is now gated by Signatera. MRD testing can help predict recurrence sooner: a 2015 study in Science found that in breast cancer, MRD offered an eight-month headstart. And if you’re a pharmaceutical company trying to develop an immunotherapy and show a measurable extension of survival time, you don’t want those people who were good with just the surgery in your control group.

Investors have already caught on to this story. As noted by The Wall Street Journal’s David Wainer last week, Natera has seen its stock price quadruple in three years and is now valued more than Illumina ($28.22 billion to $21.24 billion,) even though Illumina provides the sequencing technology underlying Signatera. The CDx designation could further accelerate adoption.

Bladder cancer accounts for only 30,000 new cases in the US each year, but if it’s any indication of what’s to come with other solid tumors, MRD testing is about to explode.

Subscribe now

Thanks to everyone who has checked out the video from my live happy hour on Friday, May 15. I lucked out as Jason Gammack, CEO of Ansa Biotechnologies was able to jump from the audience into the livestream. I’ll be hosting more of these in the future, albeit with more advanced notice. If you’d like to volunteer to join as a guest, send me a message or an email at ahan@iongenomics.bio.

Aside from the MRD regulatory news, several funding announcements caught my attention this week.

In 2021, I wrote about a new method called RABID-seq which used a modified rabies virus to track cell-to-cell interactions in the brain. The modified virus deposited barcodes in each cell it infected, showing how cells were communicating. Now, Violet Therapeutics, a company trying to use that method to find new drug targets in neurodegenerative diseases, has raised $4.8 million.

Blank Bio, a startup founded by several associates of Bo Wang’s lab at the University of Toronto, announced a partnership with Pacific Biosciences to sequence RNA to feed its AI models and a $7.2 million seed financing round.

They’re convinced that RNA data can help them predict patient prognosis and response to treatment in cancer and they’re building models to help with patient selection in clinical trials.

It’s not immediately clear whether Wang, who developed scGPT and is a cofounder of Xaira Therapeutics, is involved in Blank Bio. Blank Bio did not immediately respond.

Avant Genomics, a University of Virginia spin-out developing automated sample preparation tools for liquid biopsy, has raised more than $3 million to advance its platform.

India’s Datar Cancer Genetics said it has received FDA clearance for CellDx-Tissue, its comprehensive genomic profiling assay for solid tumors.

And early last week the Translational Genomics Research Institute announced that Nhan Tran would join as VP of Cancer Discovery and Translation and codirect its Immunology and Microbiome Division. Tran comes to TGen from the Mayo Clinic.

Elsewhere on the Internet

Illumina CEO Jacob Thaysen quietly joined President Donald Trump on a recent visit to China, joining a host of other business leaders from US finance and technology companies, according to a report from Huanjia Zhang at GenomeWeb. Thaysen’s inclusion has not been well publicized by Illumina, which has not been able to sell new instruments in the country since being placed on a government blacklist in February 2025 and has seen revenue from the country continue to fall.

In a May 19 LinkedIn post, Thaysen said “it was a privilege to join President Trump” and suggested his participation came at the invitation of the US government.

Blank Bio also raised $7.2 million in seed financing, led by Define Ventures, Leonis Capital, Nova Threshold, Ripple Ventures, SignalFire, Y Combinator, and other investors.

“This dataset will give us a richer view of patient tumours, including splice isoforms, mutations, and expression signals, helping our models learn from a more complete picture of the transcriptome,” Blank Bio CEO and Cofounder Jonathan Hsu said in a May 19 LinkedIn post. The company’s models, trained on RNA data, could help predict a patient’s disease progression and response to treatment. “By capturing more of each patient’s biology, our goal is to help clinical teams design smaller, more efficient trials that are more likely to succeed.”

Proceeds from the seed round will fund model development, expanded collaborations with pharmaceutical and diagnostic companies, and new long-read RNA-seq dataset generation.

Founded in 2025, Blank Bio has participated in the Y Combinator startup accelerator. Its cofounders also include Philip Fradkin and Ian Shi, both of whom worked in the University of Toronto lab of Bo Wang, who developed the AI foundation model scGPT and cofounded Xaira Therapeutics. Fradkin and Shi were first authors on a paper published in April in Nature Methods on an RNA-based foundation model called “Orthrus.” Whether the company is commercializing Orthrus or if Wang is involved in Blank Bio isn’t clear; the company did not immediately respond to a request for comment.

Under their collaboration, Blank Bio will generate PacBio HiFi long-read bulk RNA sequencing data from up to 100 fresh-frozen patient tumor samples across multiple cancer indications. Sequencing will be conducted at Seattle Children’s Research Institute. Blank Bio will use the resulting data to train and evaluate its models, with a focus on RNA-level signals that may improve patient stratification, biomarker discovery, and clinical interpretation. Financial and other terms of the deal were not disclosed.

The collaboration aims to elucidate obscure isoform architecture, mutational complexity, and other features of patient-specific tumor biology that are lost in gene counts provided by existing RNA datasets.

Lifespan Vision Ventures (LVV) led the round, with participation from Dementia Discovery Fund, UTEC, Ono Venture Investment, and Mass General Brigham Ventures.

Proceeds will support activities for its lead small molecule program targeting EphB3, a receptor tyrosine kinase implicated in microglia-astrocyte signaling. The company is evaluating how how glial signaling contributes to the preservation and repair of synapses in the brain, or lack thereof.

EphB3 emerged from Violet’s proprietary CONNECT platform, which maps cell-to-cell signaling interactions in the brain to identify pathways involved in brain inflammation, neurodegeneration, and synaptic loss.

CONNECT is in turn based on RABID-seq, a method developed by researchers from the Broad Institute and Harvard Medical School that traces cell-cell interactions using modified rabies virus with a barcoding scheme that is read out with single-cell sequencing. [Editor’s note: I first covered RABID-seq in 2021 for GenomeWeb.]

The Centers for Medicare & Medicaid Services’ (CMS) Molecular Diagnostics Services Program (MolDX) will cover the test for recurrence monitoring following definitive treatment with curative intent in MIBC patients. The test will be available for clinicians to order on June 1, 2026.

The coverage marks the first CMS decision for Veracyte’s whole-genome sequencing-based TrueMRD platform and the commercial launch of the company’s first minimal residual disease offering.

In morning trading on May 18, shares of Veracyte were up 6 percent at $40.80.

“Surveillance for muscle-invasive bladder cancer relies heavily on imaging right now, which has real limitations in detecting early recurrence. A whole-genome MRD test that can help identify disease recurrence, often months before imaging, gives clinicians a powerful new tool,” Matthew Galsky, Deputy Director of the Mount Sinai Tisch Cancer Center, said in a statement. “The expanding clinical evidence behind Veracyte’s MRD platform is compelling, and Medicare coverage brings this test to patients at a pivotal time when the landscape of treatment for muscle-invasive bladder cancer is rapidly changing with the development of more effective systemic therapies and expanding bladder-sparing approaches.”

Up to half of MIBC patients experience recurrence within two years of initial treatment, and current surveillance relies primarily on imaging, which may not detect recurrence until disease has substantially progressed.

VeraCyte’s TrueMRD platform analyzes patient-specific tumor variants across the entire genome in each sample, allowing longitudinal tracking of tumor clonality and molecular evolution. Monitoring disease status can guide clinicians and patients in choosing treatment options, such as selecting more aggressive therapy if disease signal returns, or avoiding riskier treatments if the signal remains low.

Clinical evidence supporting the platform includes PAGER, a prospective study published in European Urology that evaluated more than 900 blood and tissue samples from 112 MIBC patients treated with neoadjuvant chemotherapy and radical cystectomy. The TrueMRD MIBC Test detected disease recurrence a median of 131 days earlier than imaging in that study.

The test is also being used in two ongoing trials: TOMBOLA, evaluating ctDNA-guided use of adjuvant immunotherapy with checkpoint inhibitors, and NEO-BLAST, which is investigating whether MIBC patients with no remaining cancer after neoadjuvant therapy can safely opt for active surveillance rather than immediate radical cystectomy.

Last week, the FDA approved Natera’s Signatera, also a tumor-informed MRD test, as a companion diagnostic for immunotherapy in MIBC.

“This clearance reflects the consistency, rigor and clinical orientation our scientific and quality teams have brought to the development of CellDx-Tissue platform,” Dadasaheb Akolkar, director of research and innovation at DCG, said in a statement. “We look forward to extending our transformational precision oncology solutions to cancer patients across global markets.”

The regulatory action sets the company up to pursue partnerships with clinical, academic, and biopharmaceutical organizations.

CellDx-Tissue is an in vitro diagnostic test that uses targeted next-generation sequencing of DNA and RNA extracted from FFPE tumor tissue. The assay covers 517 cancer-associated genes and is designed to detect single nucleotide variants, small insertions and deletions, ERBB2 gene amplification, and gene fusions involving ALK, RET, and ROS1. The test is performed at DCG’s CAP- and CLIA-accredited laboratory in Nashik, India.

The company has previously gained FDA Breakthrough Device Designation for its liquid biopsy platforms.

I’ll definitely be doing more of these in the future, albeit with more advanced notice.

Get more from Andrew P. Han in the Substack app

Available for iOS and Android

Get the app

The approval, announced May 15, marks the first time a blood-based molecular minimal residual disease (MRD) test has been cleared as a companion diagnostic. The decision allows clinicians to use circulating tumor DNA MRD status to guide treatment decisions: initiating immunotherapy in MRD-positive patients and potentially deferring treatment in those who test negative.

The approval follows publication of the Phase III IMvigor011 trial in the New England Journal of Medicine in October 2025.

https://www.nejm.org/doi/full/10.1056/NEJMoa2511885

“Historically, we relied on imaging to tell us when cancer had returned, but that also meant millions of cancer cells were already present in the body,” Thomas Powles, principal investigator of the IMvigor011 trial and chair of the UK’s Barts Cancer Centre at St. Bartholomew’s Hospital, said in a statement. “As we saw with IMvigor011 and in several other trials, Signatera detected tumor DNA at an earlier timepoint and provided us with a significant head start to improve outcomes for patients.”

The Genentech-sponsored trial found that MIBC patients who tested positive on Signatera and received immunotherapy achieved significant improvements in disease-free survival and overall survival. MRD-negative patients who received no adjuvant therapy achieved 97 percent two-year overall survival.

How much this therapy will be prescribed, however, remains to be seen. Approximately 30,000 new MIBC cases are diagnosed annually in the U.S. and 150,000 globally. Surgery achieves long-term disease control in roughly half of patients, but identifying which patients are likely to recur has historically been difficult.

“While IMvigor011 is a practice changing trial in theory, key opinion leaders noted that the standard of care continues to evolve, and that as more patients receive neoadjuvant treatment fewer will also require adjuvant treatment after surgery,” Guggenheim Securities Analyst Subbu Nambi wrote in a May 15 note to investors. “Therefore, while [they] are positive on the IMvigor011 trial, they have been less optimistic as to how much this therapy will be used in practice.”

The news comes a week after Austin, Texas-based Natera reported first quarter revenue of $696.6 million, up 39 percent year over year from $501.8 million in Q1 2025.

Oncology testing was up more than 50 percent to 258,900 tests from 167,700 in the year ago quarter.

Natera has a growing portfolio of trials examining treatment based on MRD testing, including the DARE trial for estrogen receptor-positive, HER2-negative breast cancer and ALTAIR for colorectal cancer.