ABS 052: Investigating Cognitive Function and BMI Dynamics through a Transcriptome-Wide Association Study

Srijani Chakraborty ¹ , Rebecca Signer ² , Laura Huckins ³

¹ Fischell Department of Bioengineering, University of Maryland, College Park
² Icahn School of Medicine at Mount Sinai
³ Department of Psychiatry, Yale School of Medicine

Van Wickle (2025) Volume 1, ABS 052

Introduction: This study aims to analyze cognitive function in relation to testing for body-mass index (BMI) dynamic loci. Cognitive function refers to mental processes such as attention, memory, reasoning, and decision-making. In genome-wide association studies (GWAS), genetic variants associated with various traits and diseases are identified. We analyzed two GWASes and created two Manhattan plots for cognitive function and insomnia, where multiple genes were identified and limited genetic information was specified. However, the function and the tissue annotations of these genes remain unclear. To address these challenges, we conducted a transcriptome-wide association study (TWAS) on cognitive function traits through a statistical method called S-Predixcan. Unlike traditional association studies, TWAS does not rely on individual variables alone. In this study, we included BMI as an individual variable to improve the accuracy and robustness of the TWAS analysis. By integrating multiple data sources and leveraging the strength of TWAS, we uncovered novel insights into 2 cognitive function canonical pathways (Sulfur Metabolism Pathway and Distal Deletion Syndrome) prevalent in 13 brain tissues and their 7 associated genes (SUOX, SULT1A1, SULT1A2, TUFM, SH2B1, NFATC2IP, and LAT), which were found to be significantly related to BMI, obesity, and eating disorders. This is a novel discovery relating the KEGG Sulfur Metabolism pathway and Distal Deletion Syndrome pathway with eating disorders. Understanding the relationship between these pathways and traits will provide a stronger understanding of what factors and genes could be responsible for eating disorders, helping develop genetic tests to screen for individuals at high risk for developing these disorders. This could also bridge the funding gap between eating disorder research and other areas of neurological research, accelerating a deeper understanding of how eating disorders are affected by cognitive function.

Methods: To perform a transcriptome-wide association study, GWAS summary statistics for cognitive function and insomnia were downloaded. Cognitive function was the primary trait, with insomnia as a negative control. Summary statistics came from two studies, with cognitive data obtained via UK Biobank (Verbal Numerical Reasoning subset). Data in GRCh37 format was aligned to GRCh38 using R, and logistic vs. linear regression differences were accounted for. Manhattan plots using Bonferroni correction visualized significant loci. After alignment, S-Predixcan was run on 13 brain tissues for both traits. Windows Powershell caused issues, so analyses were ultimately executed using Command Prompt. Post-analysis, R was used to extract significant genes, annotate them, and assess overlap with BMI-related QTLs. FUMA was used to annotate S-Predixcan outputs for each tissue. Significant genes were isolated, cross-referenced, and filtered by canonical pathways. A scatterplot of p-values across pathways was generated to identify key tissue-specific genetic associations with cognitive function.

Results: Initial Manhattan plots showed 24 significant loci for cognitive function and seven for insomnia, indicating more genetic associations with cognitive function. S-Predixcan results reinforced this, revealing more Bonferroni-significant and BMI-QTL overlapping genes in cognitive function than insomnia. FUMA analysis of genes across 13 brain tissues identified two key canonical pathways: WP_16P112_DISTAL_DELETION_SYNDROME and KEGG_SULFUR_METABOLISM. The KEGG pathway (adjusted P ≥ 0.3) included SUOX, SULT1A1, and SULT1A2, while the WP pathway featured TUFM, SH2B1, NFATC2IP, and LAT. These findings suggest stronger and broader genetic involvement in cognitive function than in insomnia.

Discussion: Cognitive function showed stronger genetic overlap with BMI than insomnia, supporting a biological link between cognition and body weight. S-Predixcan and FUMA analyses highlighted two key pathways: KEGG Sulfur Metabolism (genes: SUOX, SULT1A1, SULT1A2) and WP Distal Deletion Syndrome (TUFM, SH2B1, NFATC2IP, LAT), both implicated in obesity, cognitive decline, and eating disorders. These results suggest shared genetic mechanisms across cognitive function, BMI, and disordered eating. However, reliance on European-centric GWAS data limits generalizability. Future studies must include diverse populations and focus on specific neurodegenerative diseases to refine associations. This research underscores genetics' role in brain–metabolism–behavior interactions.