Podcast recap: Heidi Rehm and Slavé Petrovski live from ASHG on breaking barriers in genomics

In a special live episode of The Genetics Podcast, recorded at the American Society of Human Genetics (ASHG) meeting, host Patrick Short was joined by two of the field’s leading voices: Dr. Heidi Rehm, Chief Genomics Officer at Massachusetts General Hospital and Co-Director of the Medical and Population Genetics Program at the Broad Institute, and Dr. Slavé Petrovski, Vice President of the Center for Genomics Research at AstraZeneca.

Together, they explored the state of genomics today, from the growing impact of large-scale data and AI to the challenges of translating genetic insights into equitable healthcare.

Where genomics is moving fastest

Heidi began by sharing one of the latest milestones from her team: a Guinness World Record for the fastest-ever clinical sequencing and diagnosis of newborns in intensive care. It took just six hours from sample to report, with sequencing completed in under four.

Beyond technological advancements, what excites her most is the cultural shift toward data sharing. After two decades of urging labs to share genetic data, Heidi says the momentum has finally flipped: “People are coming to me asking how to share, not whether to share.”

Slavé echoed that sentiment from the pharmaceutical side. Since AstraZeneca launched its company-wide genomics initiative in 2018, he’s seen a transformation in both the scale and diversity of human datasets available. The company’s multi-omics platform now integrates genomics, transcriptomics, proteomics, and imaging data to guide everything from early target discovery to clinical trial design. “We can now learn about human biology from humans at a scale that wasn’t imaginable five years ago,” he said.

Bringing genomics into real healthcare systems

Heidi described her experience transitioning from lab director to Chief Genomics Officer at Mass General. This represented a shift from developing tests to ensuring patients actually receive them. She has since built a genomic medicine unit that embeds genetic services into every hospital department, from cardiology to primary care.

Heidi shared that even patients at top hospitals still don’t get access to genomics. To address this gap, her team’s newest initiative works directly with primary care clinics to provide genetic counseling, virtual consent, and testing support. The aim is to allow underrepresented populations to access screening for hereditary cancer, carrier status, and other genetic risks without specialist referral. 

From biobanks to breakthroughs

Slavé outlined AstraZeneca’s approach to integrating genomics throughout R&D, starting with broad research consent in clinical trials, and expanding through partnerships with academic biobanks and population-scale cohorts like the UK Biobank.

He stressed the importance of data sharing beyond institutional walls, noting AstraZeneca’s open-access genetics portal is now used by researchers in more than 100 countries. He mentioned that since it would be impossible for one company or one lab to extract all the value from these datasets, they publish discoveries so the global community can push translation forward together.

Heidi agreed with this and highlighted that making the most out of data is a way to give back to patients who share it altruistically. She said, “They want it used. Our job is to make sure we’re not the blockers.”

AI and machine learning in genomics

Both guests acknowledged the explosion of AI tools in genomics, but cautioned against overstating their impact without robust datasets. Slavé emphasized the importance of data quality and having complete, well-characterized, and diverse datasets that algorithms can train on. 

Heidi described success using AI to extract structured evidence from medical literature, reducing the workload for variant curation. On the other hand, she pointed out the irony that data is often structured before publication, then locked behind text and paywalls. “If we shared structured data from the start, we wouldn’t have to use AI to reconstruct it,” she said.

Proteomics, prediction, and the next frontier

Slavé shared results from AstraZeneca’s 50,000-sample pilot in the UK Biobank, which has now scaled to 500,000 participants. Using plasma proteomics and AI modeling, his team identified protein signatures that can predict the onset of certain diseases up to 15 years before diagnosis with 90% accuracy. 

Heidi sees parallel promise for transcriptomics and methylation profiling in rare diseases, particularly for resolving undiagnosed cases. Combining these omics layers, she said, will be key to unlocking the “one in a million” conditions that no single lab or study can solve alone.

Equity and global diversity in genomics

Both guests emphasized that genomic progress will stall without greater diversity in data. Heidi discussed new efforts to expand the gnomAD database through a federated “Nomad” network, allowing countries with strict data residency laws to share variant-level summaries rather than whole genomes.

Slavé added that some of AstraZeneca’s most important discoveries now come from non-European cohorts, revealing variants entirely absent in previous datasets. “To truly understand human biology, we have to study all humans,” he said.

Gene therapies, safety, and what’s next

The conversation turned to genetic medicines, where both optimism and caution remain high. Slavé noted that every successful modality had its learning phase. “We’re still figuring out targeted delivery and long-term safety,” he said. “Once someone cracks that, it will unlock the next generation of therapies.”

Heidi underscored the need for collaboration in ultra-rare diseases, where commercial incentives are weak and patient numbers small. Her team’s “Genie” prevalence calculator helps advocacy groups quantify how common their conditions really are. This data can then be used to convince biotech partners to pursue programs once considered too rare. Both agreed that sustainable progress requires shared data, safe design, and regulatory flexibility for rare populations.

The next generation of genomics leaders

When asked what advice they’d give early-career scientists, Slavé urged curiosity and data fluency, mentioning that the biggest opportunities lie in understanding mechanisms and how targets actually affect human biology.

Heidi said she’d follow almost the same career path again, but with one addition: “If I could go back, I’d build a stronger background in data science. Everything in this field now depends on it.”

Listen to the full episode below.