Steve Ayers, PhD

Author: Steve Ayers

  • Why Genomic Results Still Land in Epic as a PDF

    A physician ordering a 500-gene panel expects the result to behave like every other lab value in Epic: discrete, filterable, plottable over time. Instead it arrives as a nine-page PDF dropped into the chart, because the interpretation platform that generated it and the EHR that displays it were never designed to agree on what a variant actually is.

    The gap isn’t a FHIR problem, though vendors will sell you that story. HL7’s genomics reporting profiles exist. Epic’s Genomics module exists. The gap is operational: someone has to map every field the interpretation platform emits to a discrete Beaker component, decide what happens when the vendor’s variant nomenclature changes between panel versions, and own the reconciliation when a reflex test updates a call that already posted to the chart.

    That’s the work nobody puts in the RFP. Vendor selection documents compare turnaround time and CAP/CLIA validation studies. They rarely ask what the discrete data mapping looks like at go-live, or who maintains it once the vendor’s annotation database updates. I’ve watched vendor migrations slip by months over exactly this, not the wet lab or the bioinformatics pipeline, but the last hundred feet between a JSON variant call and a queryable field a clinician can trend.

    If you’re standing up genomic results in Epic, or evaluating a vendor that claims to already be “integrated,” ask for three things before you sign: a sample message showing what actually posts as discrete data versus PDF-only, the mapping table between their variant classification and your flowsheet rows, and a documented process for what happens to already-posted results when a variant gets reclassified. If the vendor can’t produce the third one, budget for building it yourself, because six months post-launch a reclassification will hit and someone will ask why the chart still shows a VUS that’s now pathogenic.

    That’s the seam. It’s unglamorous, it’s rarely staffed correctly, and it’s where most clinical genomics programs actually stall.

  • The Critical Path for Clinical Genomics: Balancing Innovation and Security

    The transition of genomic science from research laboratories into routine clinical workflows is accelerating. As we generate and integrate multi-omic datasets into everyday patient care, we must ensure our informatics infrastructure matures at the exact same pace.

    Managing the AI Revolution in Genomics

    AI is adding capabilities in the area of genomics, but it must be carefully managed. Machine learning models are exceptionally useful for identifying variant patterns within massive datasets. However, these tools cannot operate as black boxes in the clinic.

    • Algorithms should augment, not replace, clinical judgment.
    • Rigorous validation of AI-derived insights is required.
    • Human expertise remains essential to interpret computational findings.

    The Imperative of Healthcare Data Security

    As sequencing becomes commonplace, the sheer volume of sensitive health data is exploding. Data security is vital to the continued advancement of the healthcare industry. Recent cybersecurity trends, particularly sophisticated ransomware attacks targeting clinical workflows, expose a critical vulnerability.

    We cannot expect continued progress in precision medicine if the underlying data architecture is compromised. Protecting immutable genetic information and patient records is a foundational priority. Robust encryption, secure data silos, and stringent access controls are non-negotiable elements of modern healthcare informatics.

    Focusing on Bench to Bedside

    Ultimately, technological and computational advancements must serve patient outcomes. Whether we are deploying new bioinformatic pipelines or refining diagnostic panels, bench to bedside is very important.

    It is not enough to develop predictive models in a vacuum. We must aggressively push these findings into real-world applications that improve patient care. The future of healthcare relies on this pragmatic integration: harnessing powerful new tools, securing our data, and remaining steadfastly focused on clinical utility.

  • Securing the Genomic Revolution: Where NGS Innovation Meets Healthcare Data Security

    The landscape of clinical genomics is evolving at a breakneck pace, but as we push the boundaries of Next-Generation Sequencing (NGS) and precision medicine, we are simultaneously navigating one of the most treacherous cybersecurity environments healthcare has ever seen. Looking at the latest developments from the past 24 hours, the intersection of genomic innovation and healthcare data security has never been more critical.

    The Rapid Maturation of NGS

    The latest trend reports indicate that NGS is definitively transitioning from a specialized research tool to a standard component of routine clinical care. We are seeing broader applications across oncology, rare disease diagnostics, and even newborn screening. The focus is shifting from simple single-gene analysis to complex polygenic models and multifactorial disease mapping.

    However, this clinical integration brings a massive data challenge. Bioinformatics and AI are becoming absolute necessities to interpret these multi-dimensional genomic datasets. The bottleneck is no longer sequencing the DNA; it is storing, analyzing, and protecting the terabytes of highly sensitive data generated per patient.

    The Escalating Threat to Healthcare Data

    Concurrently, healthcare data security reports paint a sobering picture. Ransomware remains the predominant threat, with cybercriminals increasingly deploying double-extortion tactics—stealing the data before encrypting it. What makes this particularly alarming for clinical genomics is that these attacks are specifically targeting clinical workflows and supply chain vulnerabilities, including third-party EHR hosts and cloud environments.

    The regulatory landscape is tightening in response, with stricter HIPAA Security Rule enforcement and mandates for continuous compliance and zero-trust architectures. The cost of a healthcare data breach is now astronomical, not just financially, but in the potential compromise of patient safety and the permanent exposure of immutable genetic data.

    The Informatics Imperative

    Given my professional background in genomics and informatics, I view this intersection as the defining challenge of our field. Genomics relies fundamentally on data liquidity—the ability to share, compare, and analyze vast datasets across institutions to find meaningful biological signals. Yet, security demands strict data compartmentalization, robust encryption, and rigorous access controls.

    We cannot treat genomics and IT security as siloed disciplines. The infrastructure that supports high-throughput sequencing must be designed with “security-by-default” principles. This means implementing network segmentation to isolate genomic sequencers, utilizing AI for preemptive threat detection within our bioinformatic pipelines, and rigorously vetting the security posture of our third-party analysis vendors.

    As we move toward 2026 and beyond, the success of precision medicine won’t just depend on the accuracy of our sequencers or the sophistication of our AI models. It will depend equally on our ability to build resilient, secure informatics architectures that protect the very patients we are trying to cure. The future of genomics is inextricably linked to the future of data security.