MyOme

The model is the engine of the risk modules, and the genome carries other value. MyOme's cardiometabolic and cancer results are integrated risk scores, models fitted to population data that combine a cross ancestry polygenic score with rare variants and clinical factors, and those modules do not exist without them. MyOme describes its approach as whole genome sequencing, advanced bioinformatics and AI.

The same genome also yields a curated screen of genes with actionable findings and medication response results, which are laboratory interpretation rather than model output, and MyOme's rare disease work for Natera is diagnostic sequencing. Remove the models and a genome screening service remains, which is why this is B rather than A.

Ask which results on a MyOme report come from a fitted risk model and which from variant interpretation.

Results reach an ordering clinician, and the controls around the scores are not described. MyOme tests are ordered through clinicians and returned as a report, so a person accountable for the patient decides what to do with a risk estimate.

Nothing published says who reviews a risk score before release, when a score is withheld, for example for ancestries or ages outside validation, how scores are presented against clinical thresholds, or how an updated model changes a result already issued.

Ask which conditions suppress a risk score, how scores outside validated groups are labeled, and how clinicians are told when a model update changes a patient's result.

The approach is named without version and update discipline. MyOme describes integrated risk models that combine a cross ancestry polygenic risk score, rare variants and clinical factors such as those in Tyrer-Cuzick, and its breast cancer paper names the development cohorts and the validation design. Sequencing uses Illumina technology and runs from whole genomes.

No model version is published, and nothing describes how models are recalibrated or how a customer learns that a score's method changed.

Ask for the version of each risk model on your reports, the change history, and whether previously issued scores are recalculated when a model is updated.

The model chain is in house and short. MyOme builds and runs its own integrated risk models in its own laboratory, with no outside model provider named or implied, and MyOme computes the breast cancer score it supplies to Natera's Empower customers. Illumina supplies sequencing technology and Oracle is named as a collaborator, without saying whether Oracle hosts patient data.

MyOme's notice says customer data is governed by business associate agreements and that de identified data may go to third parties, without naming processors, and no subprocessor list is published.

Ask for the list of processors that handle identifiable genomes, including any cloud host, and the terms on which de identified data reaches third parties.

Peer reviewed validation of risk prediction, and the outcome claims are still being tested. The breast cancer model was published in JCO Precision Oncology (2023): developed on 10,927 cases and 114,390 controls from UK Biobank, the Multiethnic Cohort and an African diaspora study, and validated in the Women's Health Initiative (25,284 women) and the remaining UK Biobank participants (119,187 women). Against Tyrer-Cuzick alone, the odds ratio per standard deviation for remaining lifetime risk rose from 1.31 to 1.88 in Hispanic women and from 1.29 to 1.78 in White women, area under the curve rose from 0.58 to 0.68 and from 0.57 to 0.66, and the score was well calibrated across deciles in every self reported ancestry group. Coronary artery disease models are published in npj Cardiovascular Health (2025) and the American Journal of Preventive Cardiology (2026). The authors include MyOme and Natera staff.

MyOme's headline claims are about lifelong benefit and cost savings, which these studies do not test; the prospective MyOme Proactive Health trial, enrolling from 2026, is meant to.

Ask for the trial protocol and endpoints, and for any results on screening uptake or events in customers using the scores.

A consent commitment is stated and the downstream use of de identified data is broad. MyOme's notice says it will never use protected health information for research without consent, and that it may create de identified or aggregated information for research, internal analysis, analytics and publications and make it available to third parties. No retention period for samples or genome data is stated, no de identification method is named, and nothing says whether customer genomes are used to train or recalibrate the risk models.

A whole genome is among the hardest data to de identify, which makes the third party use clause the first thing to clarify. Safety engineering for the risk models, meaning how a faulty score would be detected and handled, is not published.

Ask how long genomes and samples are kept, whether they refine MyOme's models, and what de identified data reaches third parties.

Both roles are stated, and the agreement is not published. MyOme's HIPAA Notice of Privacy Practices, effective February 2025, says MyOme uses and discloses protected health information when it acts as a covered entity, and that customer protected health information is governed by the business associate agreements it has in place with customers rather than by the notice. The notice describes patient rights, including correction, and says MyOme will never use protected health information for research without consent.

The business associate agreement offered to employers and health systems is not published, and the notice does not say which programs fall under which role.

Ask whether your program makes MyOme your business associate or a covered entity provider, and for the agreement that applies.

Nothing independent is published. MyOme displays a HIPAA compliance badge, which is a compliance claim rather than an audit of how it runs security, and no SOC 2 report, HITRUST certification, ISO/IEC 27001 certificate, penetration testing statement or trust center is published. Laboratory accreditation covers testing quality rather than information security.

The estate holds whole genomes for screening programs run for employers and health systems.

Ask for a current security attestation with its scope and date, and how genomes are protected at rest and in processing.

No device claim is made, and the scores are delivered as laboratory results. MyOme describes a CLIA certified and CAP accredited laboratory, and its tests and risk scores are reported from it on a clinician's order. No FDA clearance is claimed for any test or model, and no regulatory position for the integrated risk scores is published.

Ask how MyOme classifies its integrated risk scores for regulatory purposes and whether any model is planned for FDA submission.

Results are compared across ancestry groups for the breast cancer model, with numbers. Polygenic scores typically lose accuracy outside European ancestry, and MyOme's model was built to address that: developed on cohorts including the Multiethnic Cohort and an African diaspora study, and validated by self reported ancestry in the Women's Health Initiative and UK Biobank (JCO Precision Oncology, 2023). The paper reports the odds ratio per standard deviation and area under the curve by group, for example 1.88 and 0.68 in Hispanic women against 1.78 and 0.66 in White women in the respective cohorts, says the score was well calibrated in every self reported ancestry group, and MyOme reports twice as many high risk Black women identified as with Tyrer-Cuzick alone. Figures for Black women sit in the paper's supplement.

The coronary artery disease models carry cross ancestry work in a poster and papers whose group results are not set out on MyOme's pages.

Ask for the supplement's figures for Black and Asian women and the coronary artery disease results by ancestry.

MyOme acts as a covered entity for the results it reports, the risk score is part of that result, and its notice gives the patient a correction route against MyOme directly. The notice says patients can ask MyOme to correct information they think is incorrect or incomplete, and that MyOme will explain any refusal in writing within 60 days.

No error rate is published for the risk scores at the thresholds that change care, and no warranty, indemnity or reissue commitment attaches to a score.

Ask for the false reassurance and false alarm rates at the thresholds MyOme reports, and what MyOme does for patients when a model update changes their risk category.

Integration is offered through partners, with no record system MyOme itself documents. MyOme's own pages name a collaboration with Oracle on genomic risk prediction at global scale without describing a record integration, a marketplace listing or a live site. Its breast cancer score reaches clinicians inside Natera's Empower hereditary cancer report, which Natera orders and returns, and MyOme tests are offered inside record workflows through Nest Genomics, a separate company's product.

No integration page, standard or named health system is published by MyOme.

Ask which record systems MyOme results reach today, at which health systems, and whether risk scores arrive as discrete data rather than inside a report.

A single hosted service with location implied. MyOme sequences and analyzes in its own accredited laboratory in the United States and returns results by report, and names a collaboration with Oracle on genomic risk prediction. Its own pages do not say where genomes are stored or processed, which cloud carries the analysis, or whether any processing happens outside the United States.

Ask where genomes and risk results are hosted, in which region, and whether any processing for US customers leaves the United States.

No price is published and how the product is sold can be established. MyOme sells MyOme Atlas tests through clinicians, to employers and benefits consultants, including through a partnership with Collective Health, and to health systems, and supplies the genome behind Natera's Zenith rare disease tests. The site offers routes for patients, providers and employers without a price, a unit of charge or an insurance position.

Ask for the price per genome and per Atlas module, whether employers pay per enrolled member or per test, and whether reanalysis of an existing genome is charged.

Coverage is named with validation behind part of it. Integrated risk scores are validated across ancestries for breast cancer and coronary artery disease, with type 2 diabetes work presented as a poster. MyOme Atlas also covers medication response and a curated screen of actionable genes, and MyOme describes risk models for kidney, liver and cognitive conditions and other cancers in its December 2025 release, without published validation for those.

The tests are offered to adults through clinicians, employers and health systems for proactive screening, rather than to people with symptoms.

Ask which conditions on the report carry published validation and which are reported from models still under evaluation.