Quan Wang

RNA editing is a prevalent posttranscriptional modification characterized by single-base alterations in RNA transcripts, leading to diverse functional consequences, such as codon changes, mRNA splicing modulation, and regulation of noncoding RNAs, including their binding sites. Although next-generation sequencing has identified over 2 million adenosine-to-inosine (A-to-I) RNA editing sites in mammalian transcriptomes, the functional significance of the majority of these sites, especially those in noncoding regions, remains poorly understood. To address this gap and provide a comprehensive resource for exploring the functional impact of RNA editing in circular RNAs (circRNAs), we conducted an in-depth analysis of A-to-I editing sites in circRNAs across eight species (Homo sapiens, Mus musculus, Macaca mulatta, Gallus gallus, Rattus norvegicus, Oryctolagus cuniculus, Sus scrofa, and Danio rerio). All gathered data have been integrated into CircAI (circRNA associated with A-to-I RNA editing), the first database to combine multispecies circRNA editing data with functional predictions. CircAI offers a user-friendly platform for exploring the functional impact of RNA editing on circRNAs, including predictions of coding potential, miRNA interactions, secondary structures, and RNA editing quantitative trait loci (edQTL). By providing detailed annotations and dynamic visualization tools, CircAI serves as a pivotal resource for advancing research on the functional roles of RNA editing in circRNAs and their implications in disease.

For the first time, this study presents the complete mitochondrial genome and phylogenetic analysis of Eonemachilus longidorsalis, an endemic species in southwest China. The mitogenome is a circular molecule of 16,569 bp in length with a base composition of 30.2% A, 27.3% T, 16.6% G, and 25.9% C. The phylogenetic analysis based on the complete mitogenomes shows that E. Longidorsalis clusters with Yunnanilus jiuchiensis, Yunnanilus pleurotaenia and Eonemachilus niger. This contribution may provide a valuable framework for completely resolving phylogenetic relationships of the family Nemacheilidae in future research.

A common strategy for the functional interpretation of genome-wide association study (GWAS) findings has been the integrative analysis of GWAS and expression data. Using this strategy, many association methods (e.g., PrediXcan and FUSION) have been successful in identifying trait-associated genes via mediating effects on RNA expression. However, these approaches often ignore the effects of splicing, which can carry as much disease risk as expression. Compared to expression data, one challenge to detect associations using splicing data is the large multiple testing burden due to multidimensional splicing events within genes. Here, we introduce a multidimensional splicing gene (MSG) approach, which consists of two stages: 1) we use sparse canonical correlation analysis (sCCA) to construct latent canonical vectors (CVs) by identifying sparse linear combinations of genetic variants and splicing events that are maximally correlated with each other; and 2) we test for the association between the genetically regulated splicing CVs and the trait of interest using GWAS summary statistics. Simulations show that MSG has proper type I error control and substantial power gains over existing multidimensional expression analysis methods (i.e., S-MultiXcan, UTMOST, and sCCA+ACAT) under diverse scenarios. When applied to the Genotype-Tissue Expression Project data and GWAS summary statistics of 14 complex human traits, MSG identified on average 83%, 115%, and 223% more significant genes than sCCA+ACAT, S-MultiXcan, and UTMOST, respectively. We highlight MSG’s applications to Alzheimer’s disease, low-density lipoprotein cholesterol, and schizophrenia, and found that the majority of MSG-identified genes would have been missed from expression-based analyses. Our results demonstrate that aggregating splicing data through MSG can improve power in identifying gene-trait associations and help better understand the genetic risk of complex traits.

Gene-based rare variant association studies (RVASs) have low power due to the infrequency of rare variants and the large multiple testing burden. To correct for multiple testing, traditional false discovery rate (FDR) procedures which depend solely on P-values are often used. Recently, Independent Hypothesis Weighting (IHW) was developed to improve the detection power while maintaining FDR control by leveraging prior information for each hypothesis. Here, we present a framework to increase power of gene-based RVASs by incorporating prior information using IHW. We first build supervised machine learning models to assign each gene a prediction score that measures its disease risk, using the input of multiple biological features, fed with high-confidence risk genes and local background genes selected near GWAS significant loci as the training set. Then we use the prediction scores as covariates to prioritize RVAS results via IHW. We demonstrate the effectiveness of this framework through applications to RVASs in schizophrenia and autism spectrum disorder. We found sizeable improvements in the number of significant associations compared to traditional FDR approaches, and independent evidence supporting the relevance of the genes identified by our framework but not traditional FDR, demonstrating the potential of our framework to improve power of gene-based RVASs.

Schizophrenia (SCZ) is highly polygenic, and thousands of genes contribute to its risk. The 145 GWAS loci identified to date do not fully reveal SCZ genetic risk pathways. In this study, we explore a cost-effective strategy to increase power of inference of novel pathways, by expanding the analysis to include sub-threshold GWAS (subGWAS) loci. We identify 180 subGWAS loci (e.g., 5 x10^-8 < P ≤ 10^-6) based on SCZ summary statistics of 40,675 cases and 64,643 controls from CLOZUK and PGC datasets, and show that subGWAS loci contain substantial true genetic association signals. We merge GWAS (sigGWAS) and subGWAS loci and identify in total 304 high-confidence risk genes (HRGs) by jointly modeling the expanded set of loci. We identify dendrite development and morphogenesis (DDM, GO:0016358 and GO:0048813) as a novel category of biological processes implicated in SCZ genetic risk. SigGWAS loci fail to detect DDM, which is predominantly enriched in subGWAS loci. Further, DDM genes are significantly enriched for heritability of SCZ, as well as bipolar disorder and major depression. Genes in this functional process show cell type specificity in neurons in both fetal and adult brains, and their involvement in SCZ risk is further supported by eQTL analysis of SCZ risk alleles. We derived induced pluripotent stem cell (iPSC) lines from sporadic SCZ patients and normal controls and observe increased neurite lengths and soma sizes in patient-derived iPSC lines along multiple time points during neuronal development, further validating the genetic findings. We also find that the implicated genes are enriched in FDA-approved drug targets, suggesting a therapeutic potential for targeting the implicated biological processes for prevention and treatment. Our results showcase that expanding the analysis to include subGWAS loci is a valuable strategy for enhancing power of uncovering disease mechanisms, especially those of weak effect size, for SCZ and other complex diseases.

Studies of complex disorders benefit from integrative analyses of multiple omics data. Yet, sample mix-ups frequently occur in multi-omics studies, weakening statistical power and risking false findings. Accurately aligning sample information, genotype, and corresponding omics data is critical for integrative analyses. We developed DRAMS ( https://github.com/Yi-Jiang/DRAMS ) to Detect and Re-Align Mixed-up Samples to address the sample mix-up problem. It uses a logistic regression model followed by a modified topological sorting algorithm to identify the potential true IDs based on data relationships of multi-omics. According to tests using simulated data, the more types of omics data used or the smaller the proportion of mix-ups, the better that DRAMS performs. Applying DRAMS to real data from the PsychENCODE BrainGVEX project, we detected and corrected 201 (12.5% of total data generated) mix-ups. Of the 21 mix-ups involving errors of racial identity, DRAMS re-assigned all samples to the correct racial cluster in the 1000 Genomes project. In doing so, quantitative trait loci (QTL) (FDR<0.01) increased by an average of 1.62-fold. The use of DRAMS in multi-omics studies will strengthen statistical power of the study and improve quality of the results. Even though very limited studies have multi-omics data in place, we expect such data will increase quickly with the needs of DRAMS. Author summary: Sample mix-up happens inevitably during sample collection, processing, and data management. It leads to reduced statistical power and sometimes false findings. It is of great importance to correct mixed-up samples before conducting any downstream analyses. We developed DRAMS to detect and re-align mixed-up samples in multi-omics studies. The basic idea of DRAMS is to align the data and labels for each sample leveraging the genetic information of multi-omics data. DRAMS corrects sample IDs following a two-step strategy. At first, it estimates pairwise genetic relatedness among all the data generated from all the individuals. Because the different data generated from the same individual should share the same genetics, we can cluster all the highly related data and consider that the data from one cluster have only one potential ID. Then, we used a “majority vote” strategy to infer the potential ID for individuals in each cluster. Other information, such as match of genetics-based and reported sexes, omics priorority, etc., were also used to direct identifying the potential IDs. It has been proved that DRAMS performs very well in both simulation and PsychENCODE BrainGVEX multi-omics data.

Aberrant activation of the Hedgehog signaling pathway has been linked to the formation of numerous cancer types, including the myogenic soft tissue sarcoma, embryonal rhabdomyosarcoma (eRMS). Here, we report PCG2, a novel mouse model in which human GLI2A, a constitutive activator of Hedgehog signaling, induced undifferentiated sarcomas that were phenotypically divergent from eRMS. Rather, sarcomas arising in PCG2 mice featured some characteristics that were reminiscent of Ewing sarcoma. Even though it is widely understood that Ewing sarcoma formation is driven by EWS-ETS gene fusions, a genetically defined mouse model is not well-established. While EWS-ETS gene fusions were not present in PCG2 sarcomas, precluding their designation as Ewing sarcoma, we did find that GLI2A induced expression of known EWS-ETS gene targets essential to Ewing pathogenesis, most notably, Nkx2.2. Moreover, we found that naïve mesenchymal progenitors originate tumors in PCG2 mice. Altogether, our work provides a novel genetic mouse model, which directly connects oncogenic Hedgehog activity to the etiology of undifferentiated soft tissue sarcomas for the first time. Implications: The finding that activation of Gli2 transcription factor is sufficient to induce Ewing-like sarcomas provides a direct transformative role of the Hedgehog signaling pathway in undifferentiated soft tissue sarcoma.

Genome-wide association studies (GWAS) have identified more than 100 schizophrenia (SCZ)-associated loci, but using these findings to illuminate disease biology remains a challenge. Here we present integrative risk gene selector (iRIGS), a Bayesian framework that integrates multi-omics data and gene networks to infer risk genes in GWAS loci. By applying iRIGS to SCZ GWAS data, we predicted a set of high-confidence risk genes, most of which are not the nearest genes to the GWAS index variants. High-confidence risk genes account for a significantly enriched heritability, as estimated by stratified linkage disequilibrium score regression. Moreover, high-confidence risk genes are predominantly expressed in brain tissues, especially prenatally, and are enriched for targets of approved drugs, suggesting opportunities to reposition existing drugs for SCZ. Thus, iRIGS can leverage accumulating functional genomics and GWAS data to advance our understanding of SCZ etiology and potential therapeutics.

Given the complexity of genome regions, prioritize the functional effects of noncoding variants remains a challenge. Although several frameworks have been proposed for the evaluation of the functionality of noncoding variants, most of them used "black boxes" methods that simplify the task as the pathogenicity/benign classification problem, which ignores the distinct regulatory mechanisms of variants and leads to less desirable performance. In this study, we developed DVAR, an unsupervised framework that leverage various biochemical and evolutionary evidence to distinguish the gene regulatory categories of variants and assess their comprehensive functional impact simultaneously. DVAR performed de novo pattern discovery in high-dimensional data and identified five regulatory clusters of noncoding variants. Leveraging the new insights into the multiple functional patterns, it measures both the between-class and the within-class functional implication of the variants to achieve accurate prioritization. Compared to other two-class learning methods, it showed improved performance in identification of clinically significant variants, fine-mapped GWAS variants, eQTLs and expression-modulating variants. Moreover, it has superior performance on disease causal variants verified by genome editing (like CRISPR-Cas9), which could provide a pre-selection strategy for genome editing technologies across the whole genome. Finally, evaluated in BioVU and UK Biobank, two large-scale DNA biobanks linked to complete electronic health records (EHR), DVAR demonstrated its effectiveness in prioritizing noncoding variants associated with medical phenotypes. The C++ and Python source codes, the pre-computed DVAR-cluster labels and DVAR scores across the whole genome are available at https://www.vumc.org/cgg/dvar. Supplementary data are available at Bioinformatics online.

Lung cancer patients treated with tyrosine kinase inhibitors (TKIs) often develop resistance. More effective and safe therapeutic agents are urgently needed to overcome TKI resistance. Here, we propose a medical genetics–based approach to identify indications for over 1,000 US Food and Drug Administration–approved (FDA-approved) drugs with high accuracy. We identified a potentially novel indication for an approved antidepressant drug, sertraline, for the treatment of non–small cell lung cancer (NSCLC). We found that sertraline inhibits the viability of NSCLC cells and shows a synergy with erlotinib. Specifically, the cotreatment of sertraline and erlotinib effectively promotes autophagic flux in cells, as indicated by LC3-II accumulation and autolysosome formation. Mechanistic studies further reveal that dual treatment of sertraline and erlotinib reciprocally regulates the AMPK/mTOR pathway in NSCLC cells. The blockade of AMPK activation decreases the anticancer efficacy of either sertraline alone or the combination. Efficacy of this combination regimen is decreased by pharmacological inhibition of autophagy or genetic knockdown of ATG5 or Beclin 1 . Importantly, our results suggest that sertraline and erlotinib combination suppress tumor growth and prolong mouse survival in an orthotopic NSCLC mouse model ( P = 0.0005). In summary, our medical genetics–based approach facilitates discovery of new anticancer indications for FDA-approved drugs for the treatment of NSCLC.