Spinning convincing stories for both true and false association signals

A genome-wide association study of anti-Müllerian hormone (AMH) levels in Samoan women

Study question

Can a genome-wide association study (GWAS) and transcriptome-wide association study (TWAS) help identify genetic variation or genes associated with circulating anti-Müllerian hormone (AMH) levels in Samoan women?

Summary answer

We identified eleven genome-wide suggestive loci (strongest association signal in ARID3A 19-946163-G-C [ p = 2.32 × 10⁻⁷]) and seven transcriptome-wide significant genes ( GINS2, SENP3, USP7, TUSC3, MAFA, METTL4, NDFIP1 [all with a p < 2.50 × 10⁻⁶]) associated with circulating AMH levels in Samoan women.

What is known already

Three prior GWASs of AMH levels identified eight loci in premenopausal women of European ancestry (AMH, MCM8, TEX41 , CHECK2, CDCA7 , EIF4EBP1, BMP4 and an uncharacterized non-coding RNA gene CTB-99A3.1 ), among which the MCM8 locus was shared among all three studies.

Study design size duration

We included a sample of 1,185 women from two independently recruited samples: a family study ( n = 212; [age: 18 to 40 years]) recruited in 2002-03 from Samoa and American Samoa; and the Soifua Manuia Study ( n = 973; age: 25 to 51 years), a crosssectional population-based study recruited in 2010 from Samoa.

Participants/materials setting methods

Serum AMH levels were measured using enzyme linked immunosorbent assays (ELISA). We performed GWASs in the two participant samples using a Cox mixed-effects model to account for AMH levels below detectable limits and adjusted for centered age, centered age², polity, and kinship via kinship matrix. The summary statistics were then meta-analyzed using a fixed-effect model. We annotated the variants with p < 1 × 10⁻⁵ and calculated posterior probability of causality for prioritization. We further annotated variants using FUMA and performed colocalization and transcriptome-wide association analysis. We also assessed whether any previously reported loci were replicated in our GWAS.

Main results and the role of chance

We identified eleven novel genome-wide suggestive loci ( p < 1 × 10⁻⁵) associated with AMH levels and replicated EIF4EBP1, a previously reported AMH locus, in the GWAS. The lead variant in ARID3A , 19-946163-G-C is in high linkage disequilibrium ( r ² = 0.79) with the known age-at-menopause variant 19-950694-G-A. Nearby KISS1R is a biologically plausibility causal gene in the region; kisspeptin regulates ovarian follicle development and has been linked to AMH levels. Further investigation of the ARID3A locus is warranted.

Limitations reasons for caution

The main limitations of our study are the small sample size for a GWAS and the use of the transcription model trained on mostly European samples from the Genotype Tissue Expression (GTEx) project, which may have led to reduced power to detect genotype-expression associations. Our findings need to be validated in larger Polynesian cohorts.

Wider implications of the findings

In addition to replicating one of the eight previously discovered AMH loci, we identified new suggestive associations. It is known that the inclusion of founder populations aids in the discovery of novel loci. These findings could enhance our understanding of AMH and AMH-related reproductive phenotypes (ovarian reserve, age at menopause, premature ovarian failure, and polycystic ovary syndrome) and help build a screening approach for women at risk for these phenotypes using genetically predicted AMH levels.

Study funding/competing interests

This work was funded by NIH grants R01-HL093093 (PI: S.T.M.), R01-HL133040 (PI: R.L.M.), and T90-DE030853 (PI: C.S. Sfeir). Molecular data for the Trans-Omics in Precision Medicine (TOPMed) Program was supported by the National Heart, Lung and Blood Institute (NHLBI). The content is solely the responsibility of the authors and does not represent the official views of the National Institutes of Health.

Species-wide genomics of kākāpō provides tools to accelerate recovery

The kākāpō is a critically endangered, intensively managed, long-lived nocturnal parrot endemic to Aotearoa New Zealand. We generated and analysed whole-genome sequence data for nearly all individuals living in early 2018 (169 individuals) to generate a high-quality species-wide genetic variant callset. We leverage extensive long-term metadata to quantify genome-wide diversity of the species over time and present new approaches using probabilistic programming, combined with a phenotype dataset spanning five decades, to disentangle phenotypic variance into environmental and genetic effects while quantifying uncertainty in small populations. We find associations for growth, disease susceptibility, clutch size and egg fertility within genic regions previously shown to influence these traits in other species. Finally, we generate breeding values to predict phenotype and illustrate that active management over the past 45 years has maintained both genome-wide diversity and diversity in breeding values and, hence, evolutionary potential. We provide new pathways for informing future conservation management decisions for kākāpō, including prioritizing individuals for translocation and monitoring individuals with poor growth or high disease risk. Overall, by explicitly addressing the challenge of the small sample size, we provide a template for the inclusion of genomic data that will be transformational for species recovery efforts around the globe.

Influenza Vaccination Is Not Associated with Increased Number of Visits for Shoulder Pain

BACKGROUND

Shoulder injury from vaccination was approved for automatic compensation from the Vaccine Injury Compensation Program (VICP)-a federal government program started in 1988 to shield the manufacturers of childhood vaccines from liability. The approval was made on the basis of case reports rather than experimental evidence. This, combined with the addition of influenza vaccination to the VICP in 2005 (which broadened coverage to include adults) and other social factors, was associated with a rapid rise in the number of claims of shoulder injury from vaccination over the last decade, which now account for more than half of all claims to the VICP. Given the high prevalence of newly symptomatic sources of shoulder pain such as rotator cuff tendinopathy, combined with the high prevalence of annual influenza vaccinations, there is a substantial risk of overlap leading to the post hoc ergo propter hoc fallacy ("after this, therefore because of this") contributing to misdiagnosis and inappropriate management of patients that perceive injury from vaccination. Records of medical care after a large number of vaccinations have a good chance of detecting serious shoulder pathology, even it is uncommon, which would result in an increased prevalence of visits for shoulder problems and specific types of shoulder pathology.

QUESTIONS/PURPOSES

Is there a difference in the proportion of visits for shoulder pain within 3 months before and after vaccination among students and faculty receiving an influenza vaccination in the shoulder?

METHODS

We studied people who were vaccinated for influenza between 2009 and 2018 at a university health service. During the study period, a comprehensive billing database identified 24,206 influenza vaccinations administered to 12,870 people (median age 20 years, range 16-77; 57% women). We had 80% power to detect a 0.1% increase in the proportion of shoulder problems after vaccination compared with before vaccination. Visits with coded ICD-9 shoulder diagnoses were identified from the electronic medical record. We compared the proportion of shoulder evaluations within 3 months before and 3 months after vaccination.

RESULTS

With the numbers available, the proportion of visits for shoulder problems were not different before (1.1% [52 of 4801]) and after vaccination (1% [40 of 3977], risk ratio 1.1 [95% CI 0.8 to 1.5]; p = 0.72). Among all vaccinations, 49% (11,834 of 24,206) were preceded or followed by an appointment within 3 months before (20% [4801 of 24,206]), after (16% [3977]), or both before and after (13% [3056]) vaccine administration, and 1.4% (170) of these visits were related to a shoulder issue. The most common reason for shoulder-related appointments was atraumatic shoulder pain (79% [134 of 170]).

CONCLUSIONS

Shoulder symptoms sufficient to seek care are notably common, even among relatively young adults, and are not more common after vaccination. Although this does not rule out an important rare pathology specific to vaccination, it seems important to consider the potential harms of assuming, based largely on chronology, that persistent shoulder pain after vaccination-something expected to be common based merely on the anticipated frequency of overlap of vaccination and common shoulder problems-represents harm from vaccine.

LEVEL OF EVIDENCE

Level III, therapeutic study.

Systemic neuro-dysregulation in depression: Evidence from genome-wide association

Depression is the world's leading cause of disability. Greater understanding of the neurobiological basis of depression is necessary for developing novel treatments with improved efficacy and acceptance. Recently, major advances have been made in the search for genetic variants associated with depression which may help to elucidate etiological mechanisms. The present review has two major objectives. First, we offer a brief review of two major biological systems with strong evidence for involvement in depression pathology: neurotransmitter systems and the stress response. Secondly, we provide a synthesis of the functions of the 269 genes implicated by the most recent genome-wide meta-analysis, supporting the importance of these systems in depression and providing insights into other possible mechanisms involving neurodevelopment, neurogenesis, and neurodegeneration. Our goal is to undertake a broad, preliminary stock-taking of the most recent hypothesis-free findings and examine the weight of the evidence supporting these existing theories and highlighting novel directions. This qualitative review and accompanying gene function table provides a valuable resource and guide for basic and translational researchers, with suggestions for future mechanistic research, leveraging genetics to prioritize studies on the neurobiological processes involved in depression etiology and treatment.

Using the structure of genome data in the design of deep neural networks for predicting amyotrophic lateral sclerosis from genotype

Motivation

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease caused by aberrations in the genome. While several disease-causing variants have been identified, a major part of heritability remains unexplained. ALS is believed to have a complex genetic basis where non-additive combinations of variants constitute disease, which cannot be picked up using the linear models employed in classical genotype–phenotype association studies. Deep learning on the other hand is highly promising for identifying such complex relations. We therefore developed a deep-learning based approach for the classification of ALS patients versus healthy individuals from the Dutch cohort of the Project MinE dataset. Based on recent insight that regulatory regions harbor the majority of disease-associated variants, we employ a two-step approach: first promoter regions that are likely associated to ALS are identified, and second individuals are classified based on their genotype in the selected genomic regions. Both steps employ a deep convolutional neural network. The network architecture accounts for the structure of genome data by applying convolution only to parts of the data where this makes sense from a genomics perspective.

Results

Our approach identifies potentially ALS-associated promoter regions, and generally outperforms other classification methods. Test results support the hypothesis that non-additive combinations of variants contribute to ALS. Architectures and protocols developed are tailored toward processing population-scale, whole-genome data. We consider this a relevant first step toward deep learning assisted genotype–phenotype association in whole genome-sized data.

Availability and implementation

Our code will be available on Github, together with a synthetic dataset (https://github.com/byin-cwi/ALS-Deeplearning). The data used in this study is available to bona-fide researchers upon request.

Supplementary information

Supplementary data are available at Bioinformatics online.

Spinning convincing stories for both true and false association signals

When interpreting genome‐wide association peaks, it is common to annotate each peak by searching for genes with plausible relationships to the trait. However, “all that glitters is not gold”—one might interpret apparent patterns in the data as plausible even when the peak is a false positive. Accordingly, we sought to see how human annotators interpreted association results containing a mixture of peaks from both the original trait and a genetically uncorrelated “synthetic” trait. Two of us prepared a mix of original and synthetic peaks of three significance categories from five different scans along with relevant literature search results and then we all annotated these regions. Three annotators also scored the strength of evidence connecting each peak to the scanned trait and the likelihood of further studying that region. While annotators found original peaks to have stronger evidence (p_Bonferroni = 0.017) and higher likelihood of further study ( p_Bonferroni = 0.006) than synthetic peaks, annotators often made convincing connections between the synthetic peaks and the original trait, finding these connections 55% of the time. These results show that it is not difficult for annotators to make convincing connections between synthetic association signals and genes found in those regions.