Fabric GEM flags ANTXR2 and diagnoses an infant in five days

AI system

System: Fabric GEM artificial-intelligence variant interpretation tool with Phevor phenotype-driven re-ranking
Model: Not assessed
Release: Not assessed
Application developer: Fabric Genomics, Oakland, California
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: Clinicians submitted the infant’s genome and symptoms for analysis. They reviewed the variants that GEM ranked and confirmed an ANTXR2 variant as the cause of her condition.
Human oversight: Not assessed
Relation to AI: GEM scored and surfaced the homozygous ANTXR2 variant at 4.7. The paper says Bayes factor scores greater than 2 are considered strong support. Phevor also identified the variant but scored it 1. That is below the Phevor score of 5 or above that the paper treats as further guidance on what to prioritize. The clinicians then confirmed and interpreted the finding and made the diagnosis. The report is written by the treating clinicians and a Rady clinical genomics scientist, not by the tool's vendor.
AI contribution: GEM scored and surfaced the homozygous ANTXR2 variant at 4.7. The paper says Bayes factor scores greater than 2 are considered strong support. Phevor also identified the variant but scored it 1. That is below the Phevor score of 5 or above that the paper treats as further guidance on what to prioritize. The clinicians then confirmed and interpreted the finding and made the diagnosis. The report is written by the treating clinicians and a Rady clinical genomics scientist, not by the tool's vendor.

Benefits

Outcome identifiers: Not assessed
Benefit: An infant received a diagnosis of infantile hyaline fibromatosis within five days, and her clinicians changed her care plan.
Category: Improved Physical Health
Status: Not assessed
Beneficiary groups: People. Organizations.
Beneficiaries: An infant and the clinicians who investigated her condition.
Reach: One infant.
Distribution and access: Not assessed
Duration: Not assessed
Expected?: Yes
Costs and harms: Not assessed

Measurement

Studies: Not assessed
Study design: Not assessed
Measurement: The researchers recorded the GEM score, the interval between sample shipment and preliminary results, and the resulting changes to care.
Setting and sample: Not assessed
Measures: The GEM score and the time from sample shipment to preliminary results.
Comparison: Assumed alternative: chromosomal microarray and targeted gene panels
Results: GEM scored the causal variant at 4.7. Preliminary results returned five days after shipment. Other genetic tests were expected to take more than a month, but their eventual results are not reported.
Additional value from AI: Not assessed
Limitations: The clinicians did not report the eventual results of the slower genetic tests or whether the parents received genetic counseling. The report documents a diagnosis and care decisions, without establishing longer survival.
Denominators: Not assessed
Materials for replication: Not assessed

Money

Kind of value: Not assessed
Amount: Not assessed
Currency and prices: Not assessed
Period and unit: Not assessed
Realization: Not assessed
Recipients and payers: Not assessed
Gross or net: Not assessed
Included costs: Not assessed
Comparison: Not assessed
Method: Not assessed
Return and payback: Not assessed

People and places

Participants: Developer: Fabric Genomics, Oakland, California. Operator: Rady Children's Institute for Genomic Medicine and Robert Wood Johnson University Hospital.
Jurisdictions: Benefit: New Brunswick, New Jersey, United States. Developer: United States. Operator: United States.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Follow-up

Conditions for replication: Rapid genome sequencing, detailed symptoms, and clinicians who can review and confirm the suggested variant are needed.
Replication attempts: 1 follow-up entries. Benefit recurred: 1.
Later findings: No later development found.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Evidence

Source numbers: AIBD-0048-source-0001–0002
Citations: Numbered references below.
Language: Not assessed
Source relationships: Operator, Developer
Original evidence: Not assessed
Access and versions: Not assessed
Supported findings: Not assessed
Evidence links: Not assessed
Evidence strength: Not assessed
Missing evidence: Not assessed

Grades

Assessed benefits: People. Organizations.
Depth: People: 3. Organizations: 1.
Reasons: The diagnosis led to a meeting that coordinated feeding-tube placement and long-term care. The authors describe a disorder that often causes death before age two. They do not report that genetic counseling was delivered. Preliminary results arrived in five days, while other genetic tests were expected to take more than a month.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: The deployer's account

Entry details

Kind

Not assessed

Related entries

Both are hospital decision-support systems meant to flag a life-threatening condition in a critically ill patient; Epic's false negatives and false positives left seriously ill inpatients unrecognised by their clinicians.

Field

Clinical genomics / rare disease diagnosis. Critically ill infant undergoing rapid whole-genome sequencing.

Purpose

Rank candidate variants from rapid whole-genome sequencing against the patient's phenotype so that clinicians can reach a molecular diagnosis quickly.

Dates

Date basis unassessed: 2023 (precision unassessed). Updated: 25 October 2026.

Stage

Not assessed

Version

SHA-256: a3075cb0abfb…

References

Other identifiers: AI Incident Database: 123.
Category mappings: Not assessed
Brief: Not assessed
Revisions: 8 revisions. Latest: 25 October 2026.

Sources

Peer-reviewed study Developer

Owen MJ, Kingsmore SF, et al. (2022). An automated 13.5 hour system for scalable diagnosis and acute management guidance for genetic diseases. Nat Commun. Led from Rady Children's Institute for Genomic Medicine, with Fabric Genomics and Illumina staff among the co-authors.
Case report Deployer

Ye GX, Ontiveros E, Ivander A, Velinov M, Simotas C (2024). Autosomal Recessive Infantile Hyaline Fibromatosis Identified Using Artificial Intelligence-Assisted Rapid Whole Genome Sequencing. Cureus.