Reclassification of genetic-based risk predictions as GWAS data accumulate

Impact of returning unsolicited genomic results to nongenetic health care providers in the eMERGE III Network

PURPOSE

As genomic sequencing becomes more common, medically actionable secondary findings will increasingly be returned to health care providers (HCPs), who will be faced with managing the resulting patient care. These findings are generally unsolicited, ie, unrelated to the sequencing indication and/or ordered by another clinician.

METHODS

To understand the impact of receiving unsolicited results, we interviewed HCPs who received genomic results for patients enrolled in the Electronic Medical Records and Genomics (eMERGE) Phase III Network, which returned results on >100 actionable genes to eMERGE participants and HCPs.

RESULTS

In total, 16 HCPs across 3 eMERGE sites were interviewed about their experience of receiving a positive (likely pathogenic or pathogenic), negative, or variant of uncertain significance result for a patient enrolled in eMERGE Phase III and about managing their patient on the basis of the result. Although unsolicited, HCPs felt responsible for managing the patient's resulting medical care. HCPs indicated that clinical utility depended on the actionability of results, and whereas comfort levels varied, confidence was improved by the availability of subspecialist consults. HCPs were concerned about patient anxiety, insurability, and missing an actionable result in the electronic health record.

CONCLUSION

Our findings help inform best practices for return of unsolicited genomic screening findings in the future.

Guideline for feedback of individual genetic research findings for genomics research in Africa

As human genomics research in Africa continues to generate large amounts of data, ethical issues arise regarding how actionable genetic information is shared with research participants. The Human Heredity and Health in Africa Consortium (H3Africa) Ethics and Community Engagement Working group acknowledged the need for such guidance, identified key issues and principles relevant to genomics research in Africa and developed a practical guideline for consideration of feeding back individual genetic results of health importance in African research projects. This included a decision flowchart, providing a logical framework to assist in decision-making and planning for human genomics research projects. Although presented in the context of the H3Africa Consortium, we believe the principles described, and the decision flowchart presented here is applicable more broadly in African genomics research.

Polygenic risk scores: An overview from bench to bedside for personalised medicine

Since the first polygenic risk score (PRS) in 2007, research in this area has progressed significantly. The increasing number of SNPs that have been identified by large scale GWAS analyses has fuelled the development of a myriad of PRSs for a wide variety of diseases and, more recently, to PRSs that potentially identify differential response to specific drugs. PRSs constitute a composite genomic biomarker and potential applications for PRSs in clinical practice encompass risk prediction and disease screening, early diagnosis, prognostication, and drug stratification to improve efficacy or reduce adverse drug reactions. Nevertheless, to our knowledge, no PRSs have yet been adopted into routine clinical practice. Beyond the technical considerations of PRS development, the major challenges that face PRSs include demonstrating clinical utility and circumnavigating the implementation of novel genomic technologies at scale into stretched healthcare systems. In this review, we discuss progress in developing disease susceptibility PRSs across multiple medical specialties, development of pharmacogenomic PRSs, and future directions for the field.

Variable selection in social-environmental data: sparse regression and tree ensemble machine learning approaches

Background

Social-environmental data obtained from the US Census is an important resource for understanding health disparities, but rarely is the full dataset utilized for analysis. A barrier to incorporating the full data is a lack of solid recommendations for variable selection, with researchers often hand-selecting a few variables. Thus, we evaluated the ability of empirical machine learning approaches to identify social-environmental factors having a true association with a health outcome.

Methods

We compared several popular machine learning methods, including penalized regressions (e.g. lasso, elastic net), and tree ensemble methods. Via simulation, we assessed the methods’ ability to identify census variables truly associated with binary and continuous outcomes while minimizing false positive results (10 true associations, 1000 total variables). We applied the most promising method to the full census data (p = 14,663 variables) linked to prostate cancer registry data (n = 76,186 cases) to identify social-environmental factors associated with advanced prostate cancer.

Results

In simulations, we found that elastic net identified many true-positive variables, while lasso provided good control of false positives. Using a combined measure of accuracy, hierarchical clustering based on Spearman’s correlation with sparse group lasso regression performed the best overall. Bayesian Adaptive Regression Trees outperformed other tree ensemble methods, but not the sparse group lasso. In the full dataset, the sparse group lasso successfully identified a subset of variables, three of which replicated earlier findings.

Conclusions

This analysis demonstrated the potential of empirical machine learning approaches to identify a small subset of census variables having a true association with the outcome, and that replicate across empiric methods. Sparse clustered regression models performed best, as they identified many true positive variables while controlling false positive discoveries.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-020-01183-9.

A Bayesian method to estimate variant-induced disease penetrance

A major challenge emerging in genomic medicine is how to assess best disease risk from rare or novel variants found in disease-related genes. The expanding volume of data generated by very large phenotyping efforts coupled to DNA sequence data presents an opportunity to reinterpret genetic liability of disease risk. Here we propose a framework to estimate the probability of disease given the presence of a genetic variant conditioned on features of that variant. We refer to this as the penetrance, the fraction of all variant heterozygotes that will present with disease. We demonstrate this methodology using a well-established disease-gene pair, the cardiac sodium channel gene SCN5A and the heart arrhythmia Brugada syndrome. From a review of 756 publications, we developed a pattern mixture algorithm, based on a Bayesian Beta-Binomial model, to generate SCN5A penetrance probabilities for the Brugada syndrome conditioned on variant-specific attributes. These probabilities are determined from variant-specific features (e.g. function, structural context, and sequence conservation) and from observations of affected and unaffected heterozygotes. Variant functional perturbation and structural context prove most predictive of Brugada syndrome penetrance.

Returning a Genomic Result for an Adult-Onset Condition to the Parents of a Newborn: Insights From the BabySeq Project

The return of information from genomic sequencing in children, especially in early life, brings up complex issues around parental autonomy, the child's future autonomy, the best interest standard, and the best interests of the family. These issues are particularly important in considering the return of genomic results for adult-onset-only conditions in children. The BabySeq Project is a randomized trial used to explore the medical, behavioral, and economic impacts of integrating genomic sequencing into the care of newborns who are healthy or sick. We discuss a case in which a variant in a gene for an actionable, adult-onset-only condition was detected, highlighting the ethical issues surrounding the return of such finding in a newborn to the newborn's parents.

Big data hurdles in precision medicine and precision public health

BACKGROUND

Nowadays, trendy research in biomedical sciences juxtaposes the term 'precision' to medicine and public health with companion words like big data, data science, and deep learning. Technological advancements permit the collection and merging of large heterogeneous datasets from different sources, from genome sequences to social media posts or from electronic health records to wearables. Additionally, complex algorithms supported by high-performance computing allow one to transform these large datasets into knowledge. Despite such progress, many barriers still exist against achieving precision medicine and precision public health interventions for the benefit of the individual and the population.

MAIN BODY

The present work focuses on analyzing both the technical and societal hurdles related to the development of prediction models of health risks, diagnoses and outcomes from integrated biomedical databases. Methodological challenges that need to be addressed include improving semantics of study designs: medical record data are inherently biased, and even the most advanced deep learning's denoising autoencoders cannot overcome the bias if not handled a priori by design. Societal challenges to face include evaluation of ethically actionable risk factors at the individual and population level; for instance, usage of gender, race, or ethnicity as risk modifiers, not as biological variables, could be replaced by modifiable environmental proxies such as lifestyle and dietary habits, household income, or access to educational resources.

CONCLUSIONS

Data science for precision medicine and public health warrants an informatics-oriented formalization of the study design and interoperability throughout all levels of the knowledge inference process, from the research semantics, to model development, and ultimately to implementation.

The BabySeq project: implementing genomic sequencing in newborns

BACKGROUND

The greatest opportunity for lifelong impact of genomic sequencing is during the newborn period. The "BabySeq Project" is a randomized trial that explores the medical, behavioral, and economic impacts of integrating genomic sequencing into the care of healthy and sick newborns.

METHODS

Families of newborns are enrolled from Boston Children's Hospital and Brigham and Women's Hospital nurseries, and half are randomized to receive genomic sequencing and a report that includes monogenic disease variants, recessive carrier variants for childhood onset or actionable disorders, and pharmacogenomic variants. All families participate in a disclosure session, which includes the return of results for those in the sequencing arm. Outcomes are collected through review of medical records and surveys of parents and health care providers and include the rationale for choice of genes and variants to report; what genomic data adds to the medical management of sick and healthy babies; and the medical, behavioral, and economic impacts of integrating genomic sequencing into the care of healthy and sick newborns.

DISCUSSION

The BabySeq Project will provide empirical data about the risks, benefits and costs of newborn genomic sequencing and will inform policy decisions related to universal genomic screening of newborns.

TRIAL REGISTRATION

The study is registered in ClinicalTrials.gov Identifier: NCT02422511 . Registration date: 10 April 2015.

Functional assessment of the BMPR2 gene in lymphoblastoid cell lines from Graves' disease patients

In this study, we analysed the possible influence of the c.419-43delT BMPR2 variant in patients with Graves' disease (GD), in a molecular basis, focusing our efforts on possible alterations in the mRNA processing and synthesis. The molecular assessment of this variant in patients with GD would shed light on the association between the BMPR2 gene and the disease. The variant was detected in 18%, 55% and 10% of patients with pulmonary arterial hypertension, GD and in general population, respectively. Patients with GD fold change showed increased BMPR2 expression when matched against the controls, with a mean of 4.21 ± 1.73 (P = 0.001); BMPR2 was overexpressed in the analysed cell cycle stages. Fold change analysis of variant carriers and non-carriers showed slight overexpression and differences between phases, but none of them were statistically significant. BMPR2 expression was confirmed in the lymphoblastoid cell lines (LCLs) with a molecular weight of 115 kD, and no differences between variant carriers and non-carriers were detected. To conclude, the BMPR2 variant c.419-19delT appears in high frequency in patients with GD, and independently of its presence, BMPR2 is overexpressed in the LCLs from the GD patients tested. This increase could be paired with the described decreased expression of transforming growth factor-β1 in thyroid tissue from patients with GD.

The genomic potential of the Aspirin in Reducing Events in the Elderly and Statins in Reducing Events in the Elderly studies

Human genetic studies are continuing to increase in size and scale, but the availability of well-phenotyped longitudinal cohorts remains rare. Significant infrastructure, investment and effort are required to establish and maintain high-quality cohorts with biobanking, genetic consent and repeated clinical data measurements. Australia currently has two such cohorts established by Monash University as part of community-based clinical trials in the elderly. Both studies involve capture of demographic, mood, cognitive performance, physical function, neuroimaging, audiometry and various clinical data types over an average of 5 years. The ASPirin in Reducing Events in the Elderly (ASPREE) cohort is comprised of 16 703 Australians aged over 70 years and 2411 Americans aged over 65 years - recruited and randomised to either daily low-dose aspirin or placebo to examine the preventative benefit of aspirin on a range of clinical outcomes. The STAtins in Reducing Events in the Elderly (STAREE) study uses a similar model, and is currently recruiting 10 000 men and women aged over 70 years across Australia randomised to either low-dose statins or placebo. Both cohorts involve biobanking and consent for genetic research, with recruitment through a network of general practitioners in the community. A combination of whole-genome and targeted sequencing approaches will allow gene-phenotype relationships to be explored within the context of detailed longitudinal data. Genetic risk factors for late-onset high-burden conditions, such as cardiovascular disease and dementia will be investigated, plus research into other areas, such as healthy ageing and disease resilience will be possible due to unique phenotypes of health.

Cancer Prevention and Treatment by Wholistic Nutrition

Cancer is traditionally considered a genetic disease. It starts with a gene mutation, often caused by environmental carcinogens that are enzymatically activated to metabolites that covalently bind to DNA. If these now-damaged carcinogen-DNA adducts are not repaired before the cell replicates, they result in a mutation, which is inherited by daughter cells and their subsequent progeny. Still more mutations are added that are thought to advance cellular independence, metastasis, and drug resistance, among other characteristics typically observed for advanced cancer. The stages of initiation, promotion and progression of cancer by mutations infer irreversibility because back mutations are exceedingly rare. Thus, treatment protocols typically are designed to remove or kill cancer cells by surgery, chemotherapy, immunotherapy and/or radiotherapy. However, empirical evidence has existed to show a fundamentally different treatment option. For example, the promotion of cancer growth and development in laboratory animals initiated by a powerful mutagen/carcinogen can be repetitively turned on and off by non-mutagenic mechanisms, even completely, by modifying the consumption of protein at relevant levels of intake. Similar but less substantiated evidence also exists for other nutrients and other cancer types. This suggests that ultimate cancer development is primarily a nutrition-responsive disease rather than a genetic disease, with the understanding that nutrition is a comprehensive, wholistic biological effect that reflects the natural contents of nutrients and related substances in whole, intact food. This perspective sharply contrasts with the contemporary inference that nutrition is the summation of individual nutrients acting independently. The spelling of 'holism' with the 'w' is meant to emphasize the empirical basis for this function. The proposition that wholistic nutrition controls and even reverses disease development suggests that cancer may be treated by nutritional intervention.

Genomic sequencing in clinical practice: applications, challenges, and opportunities

The development of massively parallel sequencing (or next-generation sequencing) has facilitated a rapid implementation of genomic sequencing in clinical medicine. Genomic sequencing (GS) is now an essential tool for evaluating rare disorders, identifying therapeutic targets in neoplasms, and screening for prenatal aneuploidy. Emerging applications, such as GS for preconception carrier screening and predisposition screening in healthy individuals, are being explored in research settings and utilized by members of the public eager to incorporate genomic information into their health management. The rapid pace of adoption has created challenges for all stakeholders in clinical GS, from standardizing variant interpretation approaches in clinical molecular laboratories to ensuring that nongeneticist clinicians are prepared for new types of clinical information. Clinical GS faces a pivotal moment, as the vast potential of new quantities and types of data enable further clinical innovation and complicated implementation questions continue to be resolved.

Reclassification of prostate cancer risk using sequentially identified SNPs: Results from the REDUCE trial

BACKGROUND

Although the clinical validity of risk-associated single nucleotide polymorphisms (SNPs) for assessment of disease susceptibility has been consistently established, risk reclassification from increasing numbers of implicated risk-associated SNPs raises concern that it is premature for clinical use. Our objective is to assess the degree and impact of risk reclassification with the increasing number of SNPs.

METHODS

A total of 3239 patients from the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) trial were included. Four genetic risk scores (GRSs) were calculated based on sets of sequentially discovered prostate cancer (PCa) risk-associated SNPs (17, 34, 51, and 68 SNPs).

RESULTS

Pair-wise correlation coefficients between sets of GRSs increased as more SNPs were included in the GRS: 0.80, 0.86, and 0.95 for 17 versus 34 SNPs, 34 versus 51 SNPs, and 51 versus 68 SNPs, respectively. Using a GRS of 1.5 as a cutoff for higher versus lower risk, reclassification rates of PCa risk decreased: 14.11%, 12.04%, and 8.15% for 17 versus 34 SNPs, 34 versus 51 SNPs, and 51 versus 68 SNPs, respectively. Evolving GRSs, nevertheless, provide a tool for further refining risk assessment. When all four sequential GRSs were considered, the detection rates of PCa for men whose GRSs were consistently <1.5, reclassified, and consistently ≥1.5 were 20.8%, 29.67%, and 39.26%, respectively (Ptrend  = 1.12 × 10-8 ). In comparison, the detection rates of PCa in men with negative or positive family history were 23.75% and 31.78%, respectively.

CONCLUSIONS

Risk assessment using currently available SNPs is justified. Multiple GRS values from evolving sets of SNPs provide a valuable tool for better refining risk.

Towards precision medicine

There is great potential for genome sequencing to enhance patient care through improved diagnostic sensitivity and more precise therapeutic targeting. To maximize this potential, genomics strategies that have been developed for genetic discovery - including DNA-sequencing technologies and analysis algorithms - need to be adapted to fit clinical needs. This will require the optimization of alignment algorithms, attention to quality-coverage metrics, tailored solutions for paralogous or low-complexity areas of the genome, and the adoption of consensus standards for variant calling and interpretation. Global sharing of this more accurate genotypic and phenotypic data will accelerate the determination of causality for novel genes or variants. Thus, a deeper understanding of disease will be realized that will allow its targeting with much greater therapeutic precision.

Clinical validity and utility of genetic risk scores in prostate cancer

Current issues related to prostate cancer (PCa) clinical care (e.g., over-screening, over-diagnosis, and over-treatment of nonaggressive PCa) call for risk assessment tools that can be combined with family history (FH) to stratify disease risk among men in the general population. Since 2007, genome-wide association studies (GWASs) have identified more than 100 SNPs associated with PCa susceptibility. In this review, we discuss (1) the validity of these PCa risk-associated SNPs, individually and collectively; (2) the various methods used for measuring the cumulative effect of multiple SNPs, including genetic risk score (GRS); (3) the adequate number of SNPs needed for risk assessment; (4) reclassification of risk based on evolving numbers of SNPs used to calculate genetic risk, (5) risk assessment for men from various racial groups, and (6) the clinical utility of genetic risk assessment. In conclusion, data available to date support the clinical validity of PCa risk-associated SNPs and GRS in risk assessment among men with or without FH. PCa risk-associated SNPs are not intended for diagnostic use; rather, they should be used the same way as FH. Combining GRS and FH can significantly improve the performance of risk assessment. Improved risk assessment may have important clinical utility in targeted PCa testing. However, clinical trials are urgently needed to evaluate this clinical utility as well as the acceptance of GRS by patients and physicians.

Personal Genomic Testing for Cancer Risk: Results From the Impact of Personal Genomics Study

Purpose Significant concerns exist regarding the potential for unwarranted behavior changes and the overuse of health care resources in response to direct-to-consumer personal genomic testing (PGT). However, little is known about customers' behaviors after PGT. Methods Longitudinal surveys were given to new customers of 23andMe (Mountain View, CA) and Pathway Genomics (San Diego, CA). Survey data were linked to individual-level PGT results through a secure data transfer process. Results Of the 1,042 customers who completed baseline and 6-month surveys (response rate, 71.2%), 762 had complete cancer-related data and were analyzed. Most customers reported that learning about their genetic risk of cancers was a motivation for testing (colorectal, 88%; prostate, 95%; breast, 94%). No customers tested positive for pathogenic mutations in highly penetrant cancer susceptibility genes. A minority of individuals received elevated single nucleotide polymorphism-based PGT cancer risk estimates (colorectal, 24%; prostate, 24%; breast, 12%). At 6 months, customers who received elevated PGT cancer risk estimates were not significantly more likely to change their diet, exercise, or advanced planning behaviors or engage in cancer screening, compared with individuals at average or reduced risk. Men who received elevated PGT prostate cancer risk estimates changed their vitamin and supplement use more than those at average or reduced risk (22% v 7.6%, respectively; adjusted odds ratio, 3.41; 95% CI, 1.44 to 8.18). Predictors of 6-month behavior include baseline behavior (exercise, vitamin or supplement use, and screening), worse health status (diet and vitamin or supplement use), and older age (advanced planning, screening). Conclusion Most adults receiving elevated direct-to-consumer PGT single nucleotide polymorphism-based cancer risk estimates did not significantly change their diet, exercise, advanced care planning, or cancer screening behaviors.