henhao
Sign in
GenomicsbioRxiv
Preprint

Genomics in healthcare: GA4GH looks to 2022

Ewan BirneyiD, Jessica J. VamathevaniD, Peter GoodhandiD

Abstract

The Global Alliance for Genomics and Health (GA4GH), the standards-setting body in genomics for healthcare, aims to accelerate biomedical advancement globally. We describe the differences between healthcare- and research-driven genomics, discuss the implications of global, population-scale collections of human data for research, and outline mission-critical considerations in ethics, regulation, technology, data protection, and society. We present a crude model for estimating the rate of healthcare-funded genomes worldwide that accounts for the preparedness of each country for genomics, and infers a progression of cancer-related sequencing over time. We estimate that over 60 million patients will have their genome sequenced in a healthcare context by 2025. This represents a large technical challenge for healthcare systems, and a huge opportunity for research. We identify eight major practical, principled arguments to support the position that virtual cohorts of 100 million people or more would have tangible research benefits.

PDFDOI
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
81

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (15 Oct 2017) · CC BY 4.0

A crude country-level model projects over 60 million patients will have healthcare-funded genome sequencing by 2025; the GA4GH authors separately give eight arguments for virtual cohorts of 100 million+.

Human genome sequencing is moving from research budgets to healthcare systems, which could produce population-scale data that research can reuse only if ethics, regulation and technical standards are coordinated across countries. The authors say this shift has long been anticipated; what they add is a rough per-country estimate of the genome volumes involved and a case for federated, service-based standards set through GA4GH.

Key results

  1. SuggestedMore than 60 million patients are projected to have their genome sequenced through healthcare by 2025.

    A crude per-country model weights each country's genomic preparedness and conservatively assumes genomics reaches 70% of cancer diagnoses by 2027. It projects 47.5 million rare-disease genomes (patients plus relatives) and 83 million cancer genomes (tumour plus matched normal) by 2025 (Figure 1). · Abstract; How many genomes?; Figure legends

  2. SuggestedRare-disease diagnosis, which the authors call arguably the most successful healthcare use of genomics so far, could involve sequencing 20 million people by 2025.

    Assumes a minimum prevalence of 1% of the population (Table 2). The model uses 2.3 genomes per proband and 1% of young children affected. Because parents are also sequenced, 2% to 5% of the population could take part where programmes exist. Multiple systems report diagnostic rates of 20% to 30%. · Rare disease; How many genomes?; Figure legends; Tables

  3. SuggestedClinical cancer genomics is modelled with slower uptake but could reach over 40 million people in the developed world by 2025.

    Half of people in G20 countries will be diagnosed with a cancer event at some point in life; Table 3 gives per-country estimates. The model starts the ramp-up of healthcare-funded cancer genomes in 2022 because treatment decisions must be made within weeks and genomic results count only if they change treatment, which needs clinical trials. · Cancer; How many genomes?; Figure legends; Tables

  4. ClaimedThe authors argue on eight grounds that virtual cohorts of 100 million or more would add research value beyond today's largest cohorts of about 1 million. The grounds include rare-event association and per-base mutation-rate modelling.

    No new data. Rare diseases at 1 in 10,000 to 1 in 1,000,000 live births need at least two observations, which motivates pooling over 10 million to 500 million people. Alleles below 0.1% frequency in specific environments need 1 million to 10 million samples. 10 to 100 million genomes would approach the 1.1e-8 per-base mutation rate, compared with around 100,000 in ExAC and gnomAD. · The need for cohorts of 10 million+ humans; 1. Robust statistical association of rare and…; 3. Ability to query very low frequency alleles…; 6. Large, diverse cohorts provide a baseline…

  5. ClaimedThe authors argue that reusing healthcare genomes across borders needs federated analysis built on service-based (REST) standards, sending the analysis to where the data sit instead of pooling the data.

    This is an argument, not data. National schemes are judged unlikely to let aggregated data leave their borders, and bespoke file-based pipelines are judged incompatible with federation. Matchmaker Exchange is offered as precedent: nearly 30 matches as of October 2017, mostly between patients in different countries. · 1. Robust statistical association of rare and…; Ethical and regulatory challenges; Technical challenges; What if there were no standards?

Methods, data and reproducibility

A perspective written from within GA4GH. It contrasts healthcare-driven and research-driven genomics across four clinical categories (infectious, rare, cancer, common disease). It also uses a crude per-country model to estimate healthcare-funded rare-disease and cancer genomes over time, from country preparedness and assumed uptake rates.

  • Model projections of rare-disease genomes sequenced per OECD country over time (Table 2; 2.3 genomes per proband, 1% of young children affected) · human
  • Model projections of cancer genomes (tumour plus matched normal) per OECD country over time (Table 3; ramp-up from 2022, 70% of cancer diagnoses by 2027) · human
Methods & tools
dbGaP (mentioned), European Genome-phenome Archive (EGA) (mentioned), Matchmaker Exchange (mentioned), ExAC (mentioned), gnomAD (mentioned), REST APIs (mentioned)
Reproducibility gaps
The projection model is not released. The text does not describe how country preparedness was quantified, what the model inputs were, or how it was parameterised; only summary tables and Figure 1 are referenced.
Limitations (6)
  • The authors call the projection model crude for a complex, policy-driven area and expect some countries to adopt genomics faster or slower than predicted. · stated by the authors
  • A radical fall in sequencing cost could bring complete population sequencing within the same time frame, so the projections could shift substantially. · stated by the authors
  • Whether cohorts beyond 1 million people bring further research benefit is described as an open question; 0.5 to 1 million may be adequate. · stated by the authors
  • Population-scale healthcare data are expected to stay within national boundaries, which restricts cross-country work to federated analysis. · stated by the authors
  • The text gives no model equations, input values, sensitivity analysis or comparison with observed sequencing counts, so the 60 million figure cannot be checked. · noted in this summary
  • The case for 100-million-person cohorts is reasoned without power calculations or simulations, and it is made by authors writing for the standards body it supports. · noted in this summary

Who should read:Read this if you want the 2017 GA4GH case, with rough country-level numbers, for why healthcare-funded sequencing could create research cohorts of tens to hundreds of millions and what federated standards that would need.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISSchwartz '18Dodin '18Jacobsen '22Chen '21Daniš '24Perkins '18
58% semantic matchGenomics
Marci L.B. Schwartz, Cara Zayac McCormick, Amanda L. Lazzeri+20 more·The American Journal of Human Genetics·Aug 9, 2018· 182 citations