Category: Uncategorized

Posted on

In Conversation with Haijiao Liu, PhD

Haijiao Liu, PhD
Haijiao Liu, PhD

As a postdoctoral researcher at the University of Pennsylvania, Haijiao Liu, PhD, helped advance tumor-on-a-chip technology, a feat of bioengineering that mimics the microenvironment of malignant human tumors. Led by Dan Dongeun Huh, PhD, a Penn Engineering professor and trailblazer of organ-on-a-chip technology, Liu and his team explanted lung adenocarcinoma tumors onto the transparent chips to test their perfusion with chimeric antigen receptor (CAR) T cells. Their findings were published in October in Nature Biotechnology, with Liu as first author.

Now on paternity leave in Toronto, Liu spoke with Lindsey Leake about the implications of this work, the challenges inherent to tumor-on-a-chip studies, and his plans to launch a lab of his own this fall.

Q: Walk me through the creation of the tumor-on-a-chip. What went into its design?

Haijiao Liu: It’s essentially inspired by the need for alternative tumor models. This is speaking to the traditionally used animal tumor models and some existing in vitro tumor models, especially for the study of immunotherapies.

For example, when I started at Penn around 2018, Penn Medicine was pioneering this immunotherapy called CAR T-cell therapy, which is basically aiming to harness the patient’s own immune system, specifically the patient’s own T cells, to help fight the cancer. Penn Medicine was demonstrating huge clinical success using this CAR T therapy to treat blood cancers, such as leukemias and lymphomas. In a huge contrast to this, the solid cancer arena has seen a limited response from this new immunotherapy. So there’s this great need to study why this has not been successful, and that comes down to the consensus that the solid tumor has this really complex microenvironment.

In the category called tumor-on-a-chip, people try to control the cultural environment, the biochemical and biophysical environment of tumor cell cultures. We can use this for simple drug testing—see how the tumor growth will be affected or how effectively they can be killed. However, these existing tumor-on-chip or in vitro tumor models are still very simple. They don’t usually recreate or reproduce the complex structure of the human solid tumors that I described, like where they often include complex vessel networks.

I took the lead to address the need and the challenges of reproducing and then investigating, or probing, the dynamic interactions between those CAR T cells and human solid tumors entirely in vitro.

Q: How does the vascularization work on the chip?

Liu: It took several years to start, from the idea of building this more advanced tumor-on-a-chip technology toward proving it’s actually useful. I started by focusing on this one aspect, which is the CAR T-cell trafficking and their functions after they traffic to fight the tumors, and that will involve the recreation of the structural interface between the tumor and this complex vascular network that’s present in human tumors.

I was inspired by in vivo tumor transplantation, where traditionally, people take human tumors and then transplant them in a bulk, intact format into animal models. So my idea was, if we want to focus more on the human biology, if we want to engineer this entirely in vitro, how about we design a vascular bedding, like a miniature living model?

We basically took advantage of the self-assembly capability of human-sourced endothelial cells, combined with certain stromal fibroblasts, or stromal cells. With a bit of optimization, engineering, tweaking, then we can allow them to form capillary-like vascular networks in our engineered models.

Q: What are the advantages of recreating the tumor microenvironment in this way? That is, is the Petri dish becoming obsolete in cancer research?

Liu: The unique advantage of this way of engineering is to have a higher level of control over the structures of the tissue-tissue interface that we can build. For example, we can engineer different culture chambers. We can engineer different access windows with this model. That allows us to construct, step by step, the vascular bedding and then the tumor transplantation. Also, by forming these perfusable vessels—by the way, we can provide the infusion and flow of the CAR T cells, just like they are infused and flow in the patient—that gives us the leverage to reconstruct, probe, and then control these tissue functions in a highly precise manner.

Q: How did you and your colleagues at Penn explore CAR T-cell activity on the chip?

Liu: We first used different functional assays, like immunostaining and ELISA (enzyme-linked immunosorbent assay) assays, to characterize how the CAR T cells are doing and how they are interacting with the tumors in our engineered model. Then we disassembled this engineered tissue to extract all the cells for flow cytometry, to further characterize their functional phenotypes.

With the help of our collaborators and other people in the lab, I took advantage of this engineered model of vascular tumors interacting with CAR T cells for multi-omics analysis. For example, I was able to extract all the cells and send them for single-cell RNA seq[uencing]. We were able to look at the gene expressions of each individual cell from all the cell types that we included in this model. In this way, we have almost like a superpower to probe and read into how each cell—including the CAR Ts and tumors and the vessels—how they are responding and interacting with each other at the molecular gene levels. This is so powerful that it helped me to discover novel interactions between these parties and also new druggable targets.

The message is that through the development of this more advanced tumor-on-a-chip technology—combined with advanced multi-omics analytics and advanced computational analysis—we were able to provide this powerful in vitro technology to apply to accelerate the development of cell therapies, such as the CAR T immunotherapies for cancer, but also other complex diseases.

Q: What are the overall implications of this latest research?

Liu: With a growing understanding of human biology at the cellular and tissue levels, I think we’re seeing that our ability to engineer and design biological systems is also growing. More than ever, we have these advances in the ability to precisely construct, investigate, and then eventually control very complex tissue functions, and even organ functions. For example, our demonstrated tumor-on-a-chip technology is like presenting a miniature sandbox; we can literally see and predict the battlefield of CAR T therapy in cancer.

If we combine these advanced engineering technologies with emerging technologies in spatial multi-omics and the unprecedented productivity of the AI revolution, we will be able to accelerate the understanding of more complex human biology and extract more biological insights, and then apply that to accelerate the development of safer and more efficacious drugs and therapies, such as immunotherapies in cancer.

Q: What are the limitations of organ-on-chip technology that need to be overcome?

Liu: I think there are challenges on two fronts. The first limitation is the lack of complexity. We’re claiming that what we just published is a sufficiently complex system for us to deeply probe and understand the dynamics of CAR T tumor interactions. Still, if we’re speaking next level of translational power or potential, then we need to pursue a higher complexity that incorporates the missing but critical components from in vivo.

The other side of the coin is that if you make this engineered model more complex, you make it more challenging to reproduce or to scale up or to translate to other labs. But also, that points to an opportunity and growing room for translation, to standardize every single step, from the construction to the analysis of these engineered models, and to automate these processes as much as possible.

Q: What do you envision for your new lab?

Liu: I have a lot of things I want to do. I’m eager to establish my own team. The overarching and the unifying theme of the new lab will be to develop the next generation of in vitro complex tissue models, or I call it assembloid tissue models. Assembloid basically means there’s a stem cell-based, three-dimensional complex tissue model that intentionally incorporates different cell types, to emulate the critical tissue-tissue interactions that determine the tissue- and organ-level functions. I still need to make a big decision where the lab could be; it could be in Canada and it could also be in China.

Q: What impact might tumors-on-a-chip have on the future of precision medicine?

Liu: It’s attracting a lot of attention from biologists and clinicians who are heavily focused on using the traditional tissue models—animal models, for example, or the simple dish cultures—for their studies of interest. So the biggest impact I can foresee with our technology is that now it’s more mature. I can see it being gradually, and maybe quickly, adapted into more traditional biological labs, to help them dissect the complex biological questions they’re asking, or to accelerate the evaluation of the exciting new drugs or therapies they’re developing. Overall, I can see that accelerate this development pipeline of new drugs and therapies in precision medicine.

 

Lindsey Leake is an award-winning, independent health reporter based outside Washington, D.C. She spent 15 years as a staff journalist at outlets including Fortune, the USA TODAY Network and Sinclair Broadcast Group. She holds an MA in Science Writing from Johns Hopkins University, an MA in Journalism and Digital Storytelling from American University and a BA from Princeton University.

The post In Conversation with Haijiao Liu, PhD appeared first on Inside Precision Medicine.

Posted on

RNA Toxicity Is a Driver of Heart Disease in Muscular Dystrophy

Researchers at Baylor College of Medicine have identified a mechanism driving progressive heart disease in people with myotonic dystrophy type 1 (DM1) that operates independently of genetic repeat expansion, providing new evidence of how cardiac damage develops and when it may be reversible. The study, published in the Journal of Clinical Investigation Insight, shows that sustained expression of expanded CUG (CUGexp) repeat RNA drives cumulative cardiac injury through structural remodeling, including myocardial fibrosis, chamber dilation, and impaired contractility, while also disrupting electrical conduction pathways that regulate heart rhythm.

“Cardiac manifestations affect most DM1 patients,” said corresponding author Thomas A. Cooper, MD. “Cardiac problems are primarily electric conduction abnormalities, seen in up to 75% of adult DM1 cases, which can result in life-threatening arrhythmias accounting for 25% mortality and the second leading cause of death in DM1.”

DM1 is an autosomal dominant disorder and the most common cause of adult-onset muscular dystrophy. It is caused by an expanded CTG repeat in the DMPK gene, with affected individuals carrying between 50 and more than 4,000 repeats, compared to five to 37 in unaffected individuals. The expanded CTG repeat causes mutant transcripts that form expanded CUG (CUGexp) RNA foci and sequester muscleblind-like (MBNL) RNA-binding proteins. The result is a loss of function of MBNL.

DM1 affects multiple systems, including skeletal muscle, the central nervous system, the gastrointestinal tract, and the heart. Cardiac impairment is common in people with the disease and is often fatal. Electrical conduction abnormalities, including prolonged PR, QRS and QTc intervals, atrioventricular block and bundle branch block, occur in up to 75% of adults with DM1. Patients may also develop atrial fibrillation, tachycardia and, less commonly, ventricular arrhythmias. Structural changes such as left ventricular dysfunction, hypertrophy, dilation and fibrosis contribute to morbidity and mortality. These conditions typically worsen with age and are more severe in males.

The Baylor team used a transgenic mouse model engineered to express toxic CUG repeat RNA in the heart. Unlike human disease, the number of CTG repeats in this model remained stable over time, allowing researchers to isolate the effects of prolonged RNA toxicity without the confounding factor of repeat expansion. The researchers measured the cardiac function of the mice over a period of 14 months.

According to the researchers, the data showed that “sustained CUGexp RNA expression caused progressive cardiac enlargement, contractile dysfunction, conduction delay, myocardial fibrosis, and reduced survival, while MBNL-dependent splicing defects remained static, consistent with the stable repeat length.” These findings indicate that cardiac deterioration can occur even without increasing loss of MBNL function. This finding runs contrary to current thinking, which centers on the role of repeat expansion as the primary driver of disease progression.

In this research, the model mice developed enlarged hearts and electrical abnormalities early on. Over time, these changes progressed to cause weakened cardiac function, fibrosis and dilation of heart chambers.

The study also sought to find whether the cardiac damage could be reversed by halting expression of the toxic RNA. When RNA expression was stopped after a short duration, heart size, electrical function and structure largely returned to normal.

Yet when exposure to the toxic RNA was prolonged, the mice did not exhibit complete recovery. While the molecular defects such as abnormal RNA splicing were fully corrected, structural damage, including fibrosis and conduction abnormalities, persisted. This shows that early intervention could be one method to prevent irreversible changes such as fibrosis, which disrupts electrical signaling and increases the risk of arrhythmias.

Prior research had linked increasing CTG repeat length with DM1 severity and earlier onset, but the extent to which other mechanisms contribute to progression remained unclear. By isolating RNA toxicity, the Baylor researchers have provided evidence that chronic cellular stress, structural remodeling and potentially other RNA-mediated effects play roles in disease pathology independent of repeat expansion.

Future research will look to identify any additional mechanisms that may contribute to disease progression, and will include studying the long-term effects of RNA toxicity, changes in RNA-binding proteins, and potential links to premature cellular aging.

The post RNA Toxicity Is a Driver of Heart Disease in Muscular Dystrophy appeared first on Inside Precision Medicine.

Posted on

‘Cervix-on-a-Chip’ Brings STI Research Closer to Real Human Biology

Studying sexually transmitted infections (STIs) has long been constrained by a fundamental problem: the available models fail to fully capture the complexity of the human body. Traditional cell cultures oversimplify biology, while animal models often do not accurately reflect human infection dynamics.

Now, scientists at the University of Maryland School of Medicine and collaborators have developed the first immune-capable “cervix-on-a-chip,” a microengineered system that recreates the human cervical environment with unprecedented realism. The work, published in Science Advances, could significantly accelerate the development of new treatments and prevention strategies for STIs.

A long-standing gap in STI research

STIs remain a major global health burden. According to the World Health Organization, nearly one million new infections occur every day worldwide, with chlamydia alone accounting for roughly 129 million cases annually. In the United States, chlamydia and gonorrhea together generate an estimated $1 billion in direct medical costs each year.

Beyond their prevalence, these infections can lead to serious complications, particularly in women, including infertility, pelvic inflammatory disease, and adverse pregnancy outcomes.

Despite this, researchers have struggled to study how infections develop and progress in the human cervix, a key site of infection, under realistic conditions.

“This new model will revolutionize how scientists study STIs,” said Jacques Ravel, PhD, co-lead author of the study. “By integrating engineering, microbiology, immunology, and microbiome science, we were able to build a model that more closely reflects human biology and the complexity of the cervical microenvironment.”

Recreating the cervix in the lab

The newly developed system belongs to a class of technologies known as organ-on-a-chip models, or microphysiological systems. These platforms are designed to mimic the structure and function of human tissues using living cells and controlled physical environments.

In this case, the researchers constructed a miniature model of the cervix using human cervical epithelial cells layered on a porous membrane, with supportive tissue cells beneath. Fluids flow across both sides of the membrane, replicating the dynamic conditions found in the body.

The model also incorporates immune cells and microbial communities, allowing scientists to study how these components interact during infection.

“A key goal was to develop a complex model system that is both practical and accessible,” said Jason Gleghorn, PhD, who led the model development. “The need for this model was particularly critical for studying the vaginal microbiome, which we know plays an important role in susceptibility to STIs.”

Capturing the role of the microbiome

One of the defining features of the cervix-on-a-chip is its ability to include different types of vaginal microbiomes, something that has been difficult to replicate in previous models.

The researchers tested the system using two of the most common STIs: chlamydia (Chlamydia trachomatis) and gonorrhea (Neisseria gonorrhoeae). They found that the outcome of infection depended strongly on the type of microbiome present.

In models dominated by Lactobacillus crispatus, a bacterial species commonly associated with vaginal health, infections were significantly limited. In contrast, when less protective microbiomes were introduced, infections became more severe.

“One of the most exciting findings was that just like in women, protective microbiomes dominated by Lactobacillus crispatus limited infection in the model,” Ravel said. “In contrast, when we introduced ‘nonoptimal’ microbiomes, infections worsened.”

These results reinforce growing evidence that the vaginal microbiome plays a central role in determining susceptibility to STIs.

Toward better treatments—and prevention

Beyond improving understanding, the new model provides a practical platform for testing potential therapies.

Because it closely mimics human biology, the cervix-on-a-chip can be used to evaluate new treatments under realistic conditions. This includes not only traditional antimicrobial drugs but also emerging approaches such as probiotics and live biotherapeutics designed to restore protective microbiomes.

“This model provides a powerful new tool to develop faster, more effective, and personalized treatments,” Ravel said. “For the first time, we can simulate what happens in the human body rather than relying solely on petri dish systems or inadequate animal models.”

A platform for broader applications

The implications of the technology extend beyond the infections tested in the study. The cervix-on-a-chip could be adapted to study a wide range of pathogens, as well as broader questions about reproductive health, inflammation, and host–microbe interactions.

The researchers emphasized accessibility in the model’s design, aiming to make it usable by scientists outside of specialized bioengineering labs. This could accelerate adoption and expand its impact across the field.

A step toward more human-relevant science

The development of immune-capable organ-on-a-chip systems represents a broader shift in biomedical research toward more human-relevant experimental models. By integrating multiple components of human biology—cells, tissues, immune responses, and microbiomes—these systems offer a more accurate view of disease processes.

In the context of STIs, where subtle interactions between host and microbes can determine outcomes, such realism is particularly valuable.

As researchers continue to refine these platforms, they may help bridge the gap between laboratory studies and real-world biology—ultimately enabling earlier, more precise, and more effective interventions.

For now, the cervix-on-a-chip marks a significant step forward, providing scientists with a tool that captures the complexity of the human cervix in a way that was previously out of reach.

The post ‘Cervix-on-a-Chip’ Brings STI Research Closer to Real Human Biology appeared first on Inside Precision Medicine.

Posted on

Rates of Treatment-Related Blood Cancers Are Increasing in Japan

Research led by Osaka International Cancer Institute in Japan shows that rates of therapy-related acute myeloid leukemia, linked to treatment for an earlier cancer, are going up as the overall number of cancer survivors increases.

As reported in the journal Cancer, the researchers showed the rates of treatment-related acute myeloid leukemia (AML) almost tripled between 1990 and 2020. The share of overall AML cases taken up by these patients has also increased from 4.4% to 8.2% over a similar time period.

With increasing numbers of cancer survivors in the U.S. and elsewhere, understanding the adverse effects of cancer treatment is becoming increasingly more important. Therapy‑related AML is a rare but serious complication of cytotoxic chemotherapy, and different chemotherapy drugs carry different levels of leukemia risk. Radiotherapy can also lead to this condition, but the risk is considered lower than that of chemotherapy.

“Several population‐based studies have analyzed the long‐term trend of therapy-related AML incidence, and the results are controversial,” write lead author Kenji Kishimoto, MD, PhD, a researcher at the Osaka International Cancer Institute, and colleagues.

“A national population‐based study did not demonstrate an increase in the incidence of therapy-related AML in Denmark between 2000 and 2013. In contrast, a significant increase in therapy-related AML incidence between 1997 and 2015 was identified in a Swedish nationwide study.”

As few such studies have been carried out in Asia, Kishimoto and colleagues analyzed how such rates have changed over time and whether the initial cancers leading to therapy-related AML have changed in Japan.

The researchers identified 9841 patients with AML in the Osaka Cancer Registry, 636 (6.5%) of whom had therapy-related AML. These patients were older when diagnosed than other AML patients at a median of 69 versus 66 years. There were also slightly more women with therapy-related AML than standard AML at 45% versus 40%. The time lag between first cancer and resultant AML was between two and 11 years (median five years).

In 1990, the incidence of therapy related AML was 0.13 per 100,000 people, but this had increased to 0.36 per 100,000 people by 2020. The share of all AML cases made up by therapy-related cases almost doubled from 4.4% in 1990 to 8.2% after 2010.

The most common earlier cancers before therapy-related AML were blood cancers (23%), breast cancer (15%), colorectal cancer (12%), and gastric cancer (9%). The mix shifted over time and gastric cancer became less common as the first cancer, while breast, head‑and‑neck, and lung cancers became more prominent.

“The study provides an important step towards better understanding how the nature of therapy-related AML is changing with the increasing number of cancer survivors,” said Kishimoto in a press release.

“Findings from this study lay the foundation for further studies to elucidate the mechanism of the change in therapy-related AML epidemiology,” add the authors.

The post Rates of Treatment-Related Blood Cancers Are Increasing in Japan appeared first on Inside Precision Medicine.

Posted on

Validity of Smartphone-Based Measurement for Assessing Lower Limb Power for Sarcopenia and Frailty Discrimination: Cross-Sectional Study

Background: Increasing life expectancy has increased focus on the health-related consequences of aging, such as sarcopenia and frailty. Given the prevalence of these conditions among older individuals and the frequent resulting long-term care needs, early detection and intervention are crucial. Objective: This study aimed to validate a novel smartphone-based system measuring acceleration during the sit-to-stand movement to detect sarcopenia and frailty. Methods: Participants were 587 individuals from the Otassha study cohort who underwent health assessments in 2023, of whom 569 (96.9%) completed 2 supervised sit-to-stand trials while holding a smartphone on the lower abdomen. Sarcopenia and frailty were diagnosed using the Asian Working Group for Sarcopenia 2019 criteria and the revised Japanese Cardiovascular Health Study criteria, respectively. Peak force, rising time (T1), and stabilization time (T2) were extracted from acceleration signals, and reproducibility was examined using the intraclass correlation coefficient (ICC(2,1)). Predictive models were developed using elastic net penalized logistic regression, and model performance was evaluated using 500 bootstrap resamples. Benchmark models using age and sex, walking speed, and grip strength were also constructed for comparison. Results: Sarcopenia and frailty were identified in 16.7% (95/569) and 9% (51/569) of the participants, respectively. Peak force demonstrated excellent reliability (ICC=0.863), whereas T1 and T2 showed lower reproducibility (ICC<0.30). For sarcopenia, the smartphone model achieved a bootstrap area under the receiver operating characteristic curve (AUC) of 0.800 and an optimism-corrected AUC of 0.781 (95% CI 0.733‐0.826), outperforming walking speed (0.663) and age and sex (0.656) and ranking second only to grip strength (0.845). For frailty, the smartphone model showed moderate discrimination, with an optimism-corrected AUC of 0.659 (95% CI 0.587‐0.736), exceeding age and sex (0.604), whereas walking speed remained the strongest predictor (0.751). Conclusions: Smartphone-derived sit-to-stand acceleration provides a practical and scalable approach for screening for sarcopenia and frailty in community-dwelling older adults. While traditional indicators such as grip strength and walking speed remain the most accurate predictors, smartphone-based measurements offer meaningful complementary information and may support large-scale functional screening and early detection initiatives in superaged societies.
Posted on

Development and Evaluation of a German Suicide Prevention Website for Men: Exploratory Study

Background: Men face a substantially higher risk of suicide. Effective suicide prevention strategies for men should specifically target gender-related risk factors, such as their lower likelihood of seeking professional help. Objective: This study investigates the use and impact of a suicide prevention website for men between March 1, 2023, and December 31, 2024. The Männer Stärken website is the first suicide prevention platform for men in Germany, with the primary aim of facilitating help-seeking behavior. The development of the platform was informed by interviews with men who had attempted suicide, as well as by existing evidence on effective communication strategies for engaging men at risk. Methods: This exploratory study combines quantitative web analytics and survey data with a qualitative analysis of open-ended responses from a feedback form. Using the web analytics tool Matomo, data were collected on the number of visits to the website and the subpages they accessed. In addition, 291 anonymous feedback forms were analyzed regarding visitors’ perceptions of the website’s helpfulness and its potential to support help-seeking behavior. A further component involved an online survey (n=40) examining whether a short suicide prevention film featured on the website could increase the intention to seek help. Results: During the study period, the website recorded 29,279 visits. A majority (n=291) of the respondents reported via the feedback form that they found the website helpful (n=201, 69.1%) and believed it could encourage help-seeking behavior (59.8%). In the evaluation of the short film, a significant increase in participants’ intentions to seek help was observed in situations involving suicidal ideation and personal difficulties and when considering professional support services. This effect was not observed with regard to informal sources of support, such as friends or family. Conclusions: The data suggest that the website is being used. Among those who completed the anonymous survey form (N=291), a majority reported that the website fulfills its primary aim of providing helpful pathways to support services. The evaluation of the short film further supports this conclusion. However, certain limitations must be acknowledged: since the data were collected in a field setting, the ability to draw firm conclusions about the characteristics or representativeness of the visitor sample is limited. In addition, the sample size for the short film evaluation was small. Nevertheless, the findings point to a clear need for gender-specific suicide prevention initiatives. They indicate promising directions for the development of effective, low-threshold measures, which merit further investigation in future research.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/075608c52d64fbf781c284052cabc6f6" />

Posted on

The one piece of data that could actually shed light on your job and AI

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here.

Within Silicon Valley’s orbit, an AI-fueled jobs apocalypse is spoken about as a given. The mood is so grim that a societal impacts researcher at Anthropic, responding Wednesday to a call for more optimistic visions of AI’s future, said there might be a recession in the near term and a “breakdown of the early-career ladder.” Her less-measured colleague Dario Amodei, the company’s CEO, has called AI “a general labor substitute for humans” that could do all jobs in less than five years. And those ideas are not just coming from Anthropic, of course. 

These conversations have unsurprisingly left many workers in a panic (and are probably contributing to support for efforts to entirely pause the construction of data centers, some of which gained steam last week). The panic isn’t being helped by lawmakers, none of whom have articulated a coherent plan for what comes next.

Even economists who have cautioned that AI has not yet cut jobs and may not result in a cliff ahead are coming around to the idea that it could have a unique and unprecedented impact on how we work. 

Alex Imas, based at the University of Chicago, is one of those economists. He shared two things with me when we spoke on Friday morning: a blunt assessment that our tools for predicting what this will look like are pretty abysmal, and a “call to arms” for economists to start collecting the one type of data that could make a plan to address AI in the workforce possible at all. 

On our abysmal tools: consider the fact that any job is made up of individual tasks. One part of a real estate agent’s job, for example, is to ask clients what sort of property they want to buy. The US government chronicled thousands of these tasks in a massive catalogue first launched in 1998 and updated regularly since then. This was the data that researchers at OpenAI used in December to judge how “exposed” a job is to AI (they found a real estate agent to be 28% exposed, for example). Then in February, Anthropic used this data in its analysis of millions of Claude conversations to see which tasks people are actually using its AI to complete and where the two lists overlapped.

But knowing the AI exposure of tasks leads to an illusory understanding of how much a given job is at risk, Imas says. “Exposure alone is a completely meaningless tool for predicting displacement,” he told me.

Sure, it is illustrative in the gloomiest case—for a job in which literally every task could be done by AI with no human direction. If it costs less for an AI model to do all those tasks than what you’re paid—which is not a given, since reasoning models and agentic AI can rack up quite a bill—and it can do them well, the job likely disappears, Imas says. This is the oft-mentioned case of the elevator operator from decades ago; maybe today’s parallel is a customer service agent solely doing phone call triage. 

But for the vast majority of jobs, the case is not so simple. And the specifics matter, too: Some jobs are likely to have dark days ahead, but knowing how and when this will play out is hard to answer when only looking at exposure.

Take writing code, for example. Someone who builds premium dating apps, let’s say, might use AI coding tools to create in one day what used to take three days. That means the worker is more productive. The worker’s employer, spending the same amount of money, can now get more output. So then will the employer want more employees or fewer? 

This is the question that Imas says should keep any policymaker up at night, because the answer will change depending on the industry. And we are operating in the dark. 

In this coder’s case, these efficiencies make it possible for dating apps to lower prices. (A skeptic might expect companies to simply pocket the gains, but in a competitive market, they risk being undercut if they do.) These lower prices will always drive some increase in demand for the apps. But how much? If millions more people want it, the company might grow and ultimately hire more engineers to meet this demand. But if demand barely ticks up—maybe the people who don’t use premium dating apps still won’t want them even at a lower price—fewer coders are needed, and layoffs will happen.

Repeat this hypothetical across every job with tasks that AI can do, and you have the most pressing economic question of our time: the specifics of price elasticity, or how much demand for something changes when its price changes. And this is the second part of what Imas emphasized last week: We don’t currently have this data across the economy. But we could. 

We do have the numbers for grocery items like cereal and milk, Imas says, because the University of Chicago partners with supermarkets to get data from their price scanners. But we don’t have such figures for tutors or web developers or dietitians (all jobs found to have “exposure” to AI, by the way). Or at least not in a way that’s been widely compiled or made accessible to researchers; sometimes it’s scattered across private companies or consultancies. 

“We need, like, a Manhattan Project to collect this,” Imas says. And we don’t need it just for jobs that could obviously be affected by AI now: “Fields that are not exposed now will become exposed in the future, so you just want to track these statistics across the entire economy.”

Getting all this information would take time and money, but Imas makes the case that it’s worth it; it would give economists the first realistic look at how our AI-enabled future could unfold and give policymakers a shot at making a plan for it.

Posted on

Machine Learning and Single-Cell Technology Combined to Drive High-Performance Cell Line Development

OneCyte, which focuses on high-throughput single-cell analysis and cell line development technologies, and Kemp Proteins, which specializes in protein engineering and expression solutions, signed a strategic partnership agreement to deliver cell line development services for biopharmaceutical companies.

The collaboration brings together OneCyte’s proprietary single-cell platform for high-throughput and high-speed clone selection with Kemp Proteins’ molecular engineering capabilities, including its machine learning–driven platform, PROTiQ™.

Biopharma companies continue to face significant challenges in cell line development, including long development cycles, suboptimal yields, and high failure rates, particularly for novel and complex molecules, according to Konstantinos Tsioris, PhD, co-founder and president of OneCyte. These challenges can delay regulatory timelines and slow the progression of therapies into the clinic.

The OneCyte-Kemp partnership addresses these pain points by integrating predictive in silico design with rapid and high throughput experimental validation, say officials at both companies. As part of the workflow, amino acid sequences are evaluated using Kemp’s PROTiQ platform to assess developability risks, identify sequence liabilities, and generate structural insights.

The optimized candidates are then paired with OneCyte’s high-performance cell line development platform, which reportedly enables identification of elite clones with higher productivity.

Unlike traditional, rigid development workflows, this integrated approach is designed to adapt quickly to the evolving needs of new therapeutic modalities, notes Tsioris.

“By combining our single-cell technology with Kemp’s deep expertise in protein expression, we are confident that we can address the hardest challenges associated with new modalities, delivering faster timelines and industry-leading titers,” he continues.

“OneCyte’s class-leading single-cell technology, stacked on top of our molecular design and expression capabilities, will provide a powerful and differentiated solution for our global biopharma customers,” says Michael Keefe, CEO of Kemp Proteins.

The post Machine Learning and Single-Cell Technology Combined to Drive High-Performance Cell Line Development appeared first on GEN – Genetic Engineering and Biotechnology News.

Posted on

STAT+: A key Medicare Advantage announcement is due today

This is the online version of STAT’s weekly email newsletter Health Care Inc. Sign up here.

Well hullo! You think you can get the last word, literally, with AI? Think again. There’s always human interaction available here: bob.herman@statnews.com.

Today’s the day

By law, the 2027 Medicare Advantage payment regulation must come out today. It will set the tone for how the Trump administration wants to work with the health insurance industry: as the “new sheriff in town” or just another friendly regulator.

Continue to STAT+ to read the full story…