Guo, M.-R. | Search bioRxiv

bioRxiv2021

Brain image-derived phenotypes yield insights into causal risk of psychiatric disorders using a Mendelian randomization study

Brain imaging-derived phenotypes (IDPs) serve as underlying functional intermediates in psychiatric disorders, both of which are highly polygenic and heritable. Observational studies have elucidated the correlation between IDPs and psychiatric disorders, yet no systematic screening of IDPs to confer the causal liability in such disorders. We conducted a bidirectional two-sample Mendelian randomization framework to explore the causality of 1,901 IDPs on 10 psychiatric disorders and established the BrainMR database (http://brainmr.online/idp2psy/Index.php). We identified 17 causalities in forward MR analyses and 14 causalities in reverse MR analyses, which all rigorously examined them through various sensitivity analyses. The top significant relationship in forward was a unit lower volume of right central lateral of the thalamic nuclei, which was causally associated with an increased anorexia nervosa risk with an estimated OR of 0.52 (95% CI 0.41-0.64, P = 3.32 x 10-9). In reverse, the top significant IDP to be affected was the area of superior segment of the circular sulcus of the insula in the right hemisphere, which was increased by the reason of SCZ risk with an estimated association effect size of 0.068 (95% CI 0.046-0.090, P = 1.12 x 10-6). Overall, our study provides unique insights into causal pathways of psychiatric disorders at the imaging levels.

neuroscience↗

bioRxiv2019

Dissecting molecular regulatory mechanisms underlying noncoding susceptibility SNPs associated with 19 autoimmune diseases using multi-omics integrative analysis

The genome-wide association studies (GWAS) have identified hundreds of susceptibility loci associated with autoimmune diseases. However, over 90% of risk variants are located in the noncoding regions, leading to great challenges in deciphering the underlying causal functional variants/genes and biological mechanisms. Previous studies focused on developing new scoring method to prioritize functional/disease-relevant variants. However, they principally incorporated annotation data across all cells/tissues while omitted the cell-specific or context-specific regulation. Moreover, limited analyses were performed to dissect the detailed molecular regulatory circuits linking functional GWAS variants to disease etiology. Here we devised a new analysis frame that incorporate hundreds of immune cell-specific multi-omics data to prioritize functional noncoding susceptibility SNPs with gene targets and further dissect their downstream molecular mechanisms and clinical applications for 19 autoimmune diseases. Most prioritized SNPs have genetic associations with transcription factors (TFs) binding, histone modification or chromatin accessibility, indicating their allelic regulatory roles on target genes. Their target genes were significantly enriched in immunologically related pathways and other immunologically related functions. We also detected long-range regulation on 90.7% of target genes including 132 ones exclusively regulated by distal SNPs (eg, CD28, IL2RA), which involves several potential key TFs (eg, CTCF), suggesting the important roles of long-range chromatin interaction in autoimmune diseases. Moreover, we identified hundreds of known or predicted druggable genes, and predicted some new potential drug targets for several autoimmune diseases, including two genes (NFKB1, SH2B3) with known drug indications on other diseases, highlighting their potential drug repurposing opportunities. In summary, our analyses may provide unique resource for future functional follow-up and drug application on autoimmune diseases, which are freely available at http://fngwas.online/. Author SummaryAutoimmune diseases are groups of complex immune system disorders with high prevalence rates and high heritabilities. Previous studies have unraveled thousands of SNPs associated with different autoimmune diseases. However, it remains largely unknown on the molecular mechanisms underlying these genetic associations. Striking, over 90% of risk SNPs are located in the noncoding region. By leveraging multiple immune cell-specific multi-omics data across genomic, epigenetic, transcriptomic and 3D chromatin interaction information, we systematically analyzed the functional variants/genes and biological mechanisms underlying genetic association on 19 autoimmune diseases. We found that most functional SNPs may affect target gene expression through altering transcription factors (TFs) binding, histone modification or chromatin accessibility. Most target genes had known immunological functions. We detected prevailing long-range chromatin interaction linking distal functional SNPs to target genes. We also identified many known drug targets and predicted some new drug target genes for several autoimmune diseases, suggesting their potential clinical applications. All analysis results and tools are available online, which may provide unique resource for future functional follow-up and drug application. Our study may help reduce the gap between traditional genetic findings and biological mechanistically exploration of disease etiologies as well as clinical drug development.

immunology↗

bioRxiv2018

Modeling circRNAs expression pattern with integrated sequence and epigenetic features identifies H3K79me2 as regulators for circRNAs expression

Circular RNAs (circRNAs) are an abundant class of noncoding RNAs with widespread, cell/tissue specific pattern. Because of their involvement in the pathogenesis of multiple disease, they are receiving increasing attention. Previous work suggested that epigenetic features might be related to circRNA expression. However, current algorithms for circRNAs prediction neglect these features, leading to constant results across different cells.\n\nHere we built a machine learning framework named CIRCScan, to predict expression status and expression levels of circRNAs in various cell lines based on sequence and epigenetic features. Both expression status and expression levels can be accurately predicted by different groups of features. For expression status, the top features were similar in different cells. However, the top features for predicting expression levels were different in different cells. Noteworthy, the importance of H3K79me2 ranked high in predicting both circRNAs expression status and levels across different cells, indicating its important role in regulating circRNAs expression. Further validation experiment in K562 confirmed that knock down of H3K79me2 did result in reduction of circRNA production.\n\nOur study offers new insights into the regulation of circRNAs by incorporating epigenetic features in prediction models in different cellular contexts.

bioinformatics↗