Ancestry patterns inferred from massive RNA-seq data

Genotype prediction of 336,463 samples from public expression data

Tens of thousands of RNA-sequencing experiments comprising hundreds of thousands of individual samples have now been performed. These data represent a broad range of experimental conditions, sequencing technologies, and hypotheses under study. The Recount project has aggregated and uniformly processed hundreds of thousands of publicly available RNA-seq samples. Most of these samples only include RNA expression measurements; genotype data for these same samples would enable a wide range of analyses including variant prioritization, eQTL analysis, and studies of allele specific expression. Here, we developed a statistical model based on the existing reference and alternative read counts from the RNA-seq experiments available through Recount3 to predict genotypes at autosomal biallelic loci in coding regions. We demonstrate the accuracy of our model using large-scale studies that measured both gene expression and genotype genome-wide. We show that our predictive model is highly accurate with 99.5% overall accuracy, 99.6% major allele accuracy, and 90.4% minor allele accuracy. Our model is robust to tissue and study effects, provided the coverage is high enough. We applied this model to genotype all the samples in Recount 3 and provide the largest ready-to-use expression repository containing genotype information. We illustrate that the predicted genotype from RNA-seq data is sufficient to unravel the underlying population structure of samples in Recount3 using Principal Component Analysis.

High placental expression of FLT1, LEP, PHYHIP and IL3RA - In persons of African ancestry with severe preeclampsia

INTRODUCTION

Mortality from preeclampsia (PE) and PE-associated morbidities are 3-to 5-fold higher in persons of African ancestry than in those of Asian and European ancestries.

METHODS

To elucidate placental contribution to worse PE outcomes in African ancestry pregnancies, we performed bulk RNA sequencing on 50 placentas from persons with severe PE (sPE) of African (n = 9), Asian (n = 18) and European (n = 23) ancestries and 73 normotensive controls of African (n = 10), Asian (n = 15) and European (n = 48) ancestries.

RESULTS

Previously described canonical preeclampsia genes, involved in metabolism and hypoxia/angiogenesis including: LEP, HK2, FSTL3, FLT1, ENG, TMEM45A, ARHGEF4 and HTRA1 were upregulated sPE versus normotensive placentas across ancestries. LTF, NPR3 and PHYHIP were higher in African vs. Asian ancestry sPE placentas. Allograft rejection/adaptive immune response genes were upregulated in placentas from African but not in Asian or European ancestry sPE patients; IL3RA was of particular interest because the patient with the highest placental IL3RA expression, a person of African ancestry with sPE, developed postpartum cardiomyopathy, and was the only patient out of 123, that developed this condition. Interestingly, the sPE patients with the highest IL3RA expression among persons of Asian and European ancestries developed unexplained tachycardia peripartum, necessitating echocardiography in the European ancestry patient. The association between elevated placental IL3RA levels and unexplained tachycardia or peripartum cardiomyopathy was found to be significant in the 50 sPE patients (p = .0005).

DISCUSSION

High placental upregulation of both canonical preeclampsia and allograft rejection/adaptive immune response genes may contribute to worse PE outcomes in African ancestry sPE patients.

Direct inference and control of genetic population structure from RNA sequencing data

RNAseq data can be used to infer genetic variants, yet its use for estimating genetic population structure remains underexplored. Here, we construct a freely available computational tool (RGStraP) to estimate RNAseq-based genetic principal components (RG-PCs) and assess whether RG-PCs can be used to control for population structure in gene expression analyses. Using whole blood samples from understudied Nepalese populations and the Geuvadis study, we show that RG-PCs had comparable results to paired array-based genotypes, with high genotype concordance and high correlations of genetic principal components, capturing subpopulations within the dataset. In differential gene expression analysis, we found that inclusion of RG-PCs as covariates reduced test statistic inflation. Our paper demonstrates that genetic population structure can be directly inferred and controlled for using RNAseq data, thus facilitating improved retrospective and future analyses of transcriptomic data.

Equitable machine learning counteracts ancestral bias in precision medicine, improving outcomes for all

Gold standard genomic datasets severely under-represent non-European populations, leading to inequities and a limited understanding of human disease [1-8]. Therapeutics and outcomes remain hidden because we lack insights that we could gain from analyzing ancestry-unbiased genomic data. To address this significant gap, we present PhyloFrame, the first-ever machine learning method for equitable genomic precision medicine. PhyloFrame corrects for ancestral bias by integrating big data tissue-specific functional interaction networks, global population variation data, and disease-relevant transcriptomic data. Application of PhyloFrame to breast, thyroid, and uterine cancers shows marked improvements in predictive power across all ancestries, less model overfitting, and a higher likelihood of identifying known cancer-related genes. The ability to provide accurate predictions for underrepresented groups, in particular, is substantially increased. These results demonstrate how AI can mitigate ancestral bias in training data and contribute to equitable representation in medical research.

A multi-tissue study of immune gene expression profiling highlights the key role of the nasal epithelium in COVID-19 severity

Coronavirus Disease-19 (COVID-19) symptoms range from mild to severe illness; the cause for this differential response to infection remains unknown. Unravelling the immune mechanisms acting at different levels of the colonization process might be key to understand these differences. We carried out a multi-tissue (nasal, buccal and blood; n = 156) gene expression analysis of immune-related genes from patients affected by different COVID-19 severities, and healthy controls through the nCounter technology. Mild and asymptomatic cases showed a powerful innate antiviral response in nasal epithelium, characterized by activation of interferon (IFN) pathway and downstream cascades, successfully controlling the infection at local level. In contrast, weak macrophage/monocyte driven innate antiviral response and lack of IFN signalling activity were present in severe cases. Consequently, oral mucosa from severe patients showed signals of viral activity, cell arresting and viral dissemination to the lower respiratory tract, which ultimately could explain the exacerbated innate immune response and impaired adaptative immune responses observed at systemic level. Results from saliva transcriptome suggest that the buccal cavity might play a key role in SARS-CoV-2 infection and dissemination in patients with worse prognosis. Co-expression network analysis adds further support to these findings, by detecting modules specifically correlated with severity involved in the abovementioned biological routes; this analysis also provides new candidate genes that might be tested as biomarkers in future studies. We also found tissue specific severity-related signatures mainly represented by genes involved in the innate immune system and cytokine/chemokine signalling. Local immune response could be key to determine the course of the systemic response and thus COVID-19 severity. Our findings provide a framework to investigate severity host gene biomarkers and pathways that might be relevant to diagnosis, prognosis, and therapy.

Clinical implementation of RNA sequencing for Mendelian disease diagnostics

BACKGROUND

Lack of functional evidence hampers variant interpretation, leaving a large proportion of individuals with a suspected Mendelian disorder without genetic diagnosis after whole genome or whole exome sequencing (WES). Research studies advocate to further sequence transcriptomes to directly and systematically probe gene expression defects. However, collection of additional biopsies and establishment of lab workflows, analytical pipelines, and defined concepts in clinical interpretation of aberrant gene expression are still needed for adopting RNA sequencing (RNA-seq) in routine diagnostics.

METHODS

We implemented an automated RNA-seq protocol and a computational workflow with which we analyzed skin fibroblasts of 303 individuals with a suspected mitochondrial disease that previously underwent WES. We also assessed through simulations how aberrant expression and mono-allelic expression tests depend on RNA-seq coverage.

RESULTS

We detected on average 12,500 genes per sample including around 60% of all disease genes-a coverage substantially higher than with whole blood, supporting the use of skin biopsies. We prioritized genes demonstrating aberrant expression, aberrant splicing, or mono-allelic expression. The pipeline required less than 1 week from sample preparation to result reporting and provided a median of eight disease-associated genes per patient for inspection. A genetic diagnosis was established for 16% of the 205 WES-inconclusive cases. Detection of aberrant expression was a major contributor to diagnosis including instances of 50% reduction, which, together with mono-allelic expression, allowed for the diagnosis of dominant disorders caused by haploinsufficiency. Moreover, calling aberrant splicing and variants from RNA-seq data enabled detecting and validating splice-disrupting variants, of which the majority fell outside WES-covered regions.

CONCLUSION

Together, these results show that streamlined experimental and computational processes can accelerate the implementation of RNA-seq in routine diagnostics.

African Americans and European Americans exhibit distinct gene expression patterns across tissues and tumors associated with immunologic functions and environmental exposures

The COVID-19 pandemic has affected African American populations disproportionately with respect to prevalence, and mortality. Expression profiles represent snapshots of combined genetic, socio-environmental (including socioeconomic and environmental factors), and physiological effects on the molecular phenotype. As such, they have potential to improve biological understanding of differences among populations, and provide therapeutic biomarkers and environmental mitigation strategies. Here, we undertook a large-scale assessment of patterns of gene expression between African Americans and European Americans, mining RNA-Seq data from 25 non-diseased and diseased (tumor) tissue-types. We observed the widespread enrichment of pathways implicated in COVID-19 and integral to inflammation and reactive oxygen stress. Chemokine CCL3L3 expression is up-regulated in African Americans. GSTM1, encoding a glutathione S-transferase that metabolizes reactive oxygen species and xenobiotics, is upregulated. The little-studied F8A2 gene is up to 40-fold more highly expressed in African Americans; F8A2 encodes HAP40 protein, which mediates endosome movement, potentially altering the cellular response to SARS-CoV-2. African American expression signatures, superimposed on single cell-RNA reference data, reveal increased number or activity of esophageal glandular cells and lung ACE2-positive basal keratinocytes. Our findings establish basal prognostic signatures that can be used to refine approaches to minimize risk of severe infection and improve precision treatment of COVID-19 for African Americans. To enable dissection of causes of divergent molecular phenotypes, we advocate routine inclusion of metadata on genomic and socio-environmental factors for human RNA-sequencing studies.

Transcriptomic analysis of endometrial receptivity for a genomic diagnostics model of Chinese women

OBJECTIVE

To define the transcriptomic signature with respect to human endometrial receptivity in Chinese women by next-generation sequencing and to develop a more refined and customized bioinformatic predictive method for endometrial dating in Chinese women.

DESIGN

Randomized.

SETTING

A tertiary hospital-based reproductive medicine center.

PATIENT(S)

Ninety healthy, fertile Chinese women.

INTERVENTION(S)

Human endometrial biopsies.

MAIN OUTCOME MEASURE(S)

Gene expression of endometrial biopsies.

RESULT(S)

Ninety endometrial samples from healthy Chinese women during their menstrual cycles-including prereceptive (luteinizing hormone [LH] + 3 days/LH + 5 days), receptive (LH + 7 days), and post-receptive (LH + 9 days) phases-were subjected to transcriptomic analysis using messenger RNA (mRNA)-enriched RNA-Seq. Feature genes were obtained and used to train the predictor for endometrial dating, with 63 samples for the training set and 27 samples for the validation set. Differentially expressed genes (DEGs) were identified by comparing samples from different phases of the menstrual cycle. Based on the transcriptomic feature genes, we constructed a bioinformatic predictor for endometrial dating. The accuracy on assessment of the endometrium on days LH + 3, LH + 5, LH + 7, and LH + 9 was 100% in the training set and 85.19% in the validation set.

CONCLUSION(S)

Our transcriptomic profiling method can be used to monitor the window of implantation with regard to the endometrium in the Chinese population. This method potentially provides an evaluation of endometrial status, and can be used to predict a personal window of implantation by reproductive medicine clinicians.

Host Transcriptomic Response Following Administration of Rotavirus Vaccine in Infants' Mimics Wild Type Infection

Background

Rotavirus (RV) is an enteric pathogen that has devastating impact on childhood morbidity and mortality worldwide. The immunologic mechanism underlying the protection achieved after RV vaccination is not yet fully understood.

Methods

We compared the transcriptome of children affected by community-acquired RV infection and children immunized with a live attenuated RV vaccine (RotaTeq®).

Results

RV vaccination mimics the wild type infection causing similar changes in children's transcriptome, including transcripts associated with cell cycle, diarrhea, nausea, vomiting, intussusception, and abnormal morphology of midgut. A machine learning approach allowed to detect a combination of nine-transcripts that differentiates vaccinated from convalescent-naturally infected children (AUC: 90%; 95%CI: 70-100) and distinguishes between acute-infected and healthy control children (in both cases, AUC: 100%; 95%CI: 100-100). We identified a miRNA hsa-mir-149 that seems to play a role in the host defense against viral pathogens and may have an antiviral role.

Discussion

Our findings might shed further light in the understanding of RV infection, its functional link to intussusception causes, as well as guide development of antiviral treatments and safer and more effective vaccines. The nine-transcript signature may constitute a marker of vaccine protection and helps to differentiate vaccinated from naturally infected or susceptible children.

Human Alveolar and Splenic Macrophage Populations Display a Distinct Transcriptomic Response to Infection With Mycobacterium tuberculosis

Mycobacterium tuberculosis (Mtb) infects alveolar macrophages (AMs), causing pulmonary tuberculosis (PTB), the most common form of the disease. Less frequently, Mtb is disseminated to many other organs and tissues, resulting in different extrapulmonary forms of TB. Nevertheless, very few studies have addressed the global mRNA response of human AMs, particularly from humans with the active form of the disease. Strikingly, almost no studies have addressed the response of human extrapulmonary macrophages to Mtb infection. In this pilot study, using microarray technology, we examined the transcriptomic ex vivo response of AMs from PTB patients (AMTBs) and AMs from control subjects (AMCTs) infected with two clinical isolates of Mtb. Furthermore, we also studied the infection response of human splenic macrophages (SMs) to Mtb isolates, as a model for extrapulmonary infection, and compared the transcriptomic response between AMs and SMs. Our results showed a striking difference in global mRNA profiles in response to infection between AMs and SMs, implicating a tissue-specific macrophage response to Mtb.

RNA-Seq Data-Mining Allows the Discovery of Two Long Non-Coding RNA Biomarkers of Viral Infection in Humans

There is a growing interest in unraveling gene expression mechanisms leading to viral host invasion and infection progression. Current findings reveal that long non-coding RNAs (lncRNAs) are implicated in the regulation of the immune system by influencing gene expression through a wide range of mechanisms. By mining whole-transcriptome shotgun sequencing (RNA-seq) data using machine learning approaches, we detected two lncRNAs (ENSG00000254680 and ENSG00000273149) that are downregulated in a wide range of viral infections and different cell types, including blood monocluclear cells, umbilical vein endothelial cells, and dermal fibroblasts. The efficiency of these two lncRNAs was positively validated in different viral phenotypic scenarios. These two lncRNAs showed a strong downregulation in virus-infected patients when compared to healthy control transcriptomes, indicating that these biomarkers are promising targets for infection diagnosis. To the best of our knowledge, this is the very first study using host lncRNAs biomarkers for the diagnosis of human viral infections.

A Meta-Analysis of Multiple Whole Blood Gene Expression Data Unveils a Diagnostic Host-Response Transcript Signature for Respiratory Syncytial Virus

Respiratory syncytial virus (RSV) is one of the major causes of acute lower respiratory tract infection worldwide. The absence of a commercial vaccine and the limited success of current therapeutic strategies against RSV make further research necessary. We used a multi-cohort analysis approach to investigate host transcriptomic biomarkers and shed further light on the molecular mechanism underlying RSV-host interactions. We meta-analyzed seven transcriptome microarray studies from the public Gene Expression Omnibus (GEO) repository containing a total of 922 samples, including RSV, healthy controls, coronaviruses, enteroviruses, influenzas, rhinoviruses, and coinfections, from both adult and pediatric patients. We identified > 1500 genes differentially expressed when comparing the transcriptomes of RSV-infected patients against healthy controls. Functional enrichment analysis showed several pathways significantly altered, including immunologic response mediated by RSV infection, pattern recognition receptors, cell cycle, and olfactory signaling. In addition, we identified a minimal 17-transcript host signature specific for RSV infection by comparing transcriptomic profiles against other respiratory viruses. These multi-genic signatures might help to investigate future drug targets against RSV infection.