Exploration into biomarker potential of region-specific brain gene co-expression networks

Abstract

The human brain is a complex organ that consists of several regions each with a unique gene expression pattern. Our intent in this study was to construct a gene co-expression network (GCN) for the normal brain using RNA expression profiles from the Genotype-Tissue Expression (GTEx) project. The brain GCN contains gene correlation relationships that are broadly present in the brain or specific to thirteen brain regions, which we later combined into six overarching brain mini-GCNs based on the brain’s structure. Using the expression profiles of brain region-specific GCN edges, we determined how well the brain region samples could be discriminated from each other, visually with t-SNE plots or quantitatively with the Gene Oracle deep learning classifier. Next, we tested these gene sets on their relevance to human tumors of brain and non-brain origin. Interestingly, we found that genes in the six brain mini-GCNs showed markedly higher mutation rates in tumors relative to matched sets of random genes. Further, we found that cortex genes subdivided Head and Neck Squamous Cell Carcinoma (HNSC) tumors and Pheochromocytoma and Paraganglioma (PCPG) tumors into distinct groups. The brain GCN and mini-GCNs are useful resources for the classification of brain regions and identification of biomarker genes for brain related phenotypes.

Human endogenous retroviruses of the HERV-K (HML-2) family are expressed in the brain of healthy individuals and modify the composition of the brain-infiltrating immune cells

“…Our results indicate an elevated HERV-K (HML-2) provirus expression in the cerebellum and cerebellar hemisphere, corroborating the findings of a study by Burn et al [ 33 ]. The cerebellum and cerebellar hemispheres have also been reported to have gene expression patterns that are distinct from all other brain regions [ 40 ]. Differences in transcriptomes partly reflect the differences in epigenetic regulation, which would also affect the expression of HERV proviruses.…”

Addressing noise in co-expression network construction

“…The third is a subset of the rice dataset that contains only 141 samples that underwent drought stress. The fourth is a large GTEx [ 65 ] normal human brain tissue dataset consisting of the expression pattern of 56 202 genes across 1671 samples and 13 brain tissues [ 50 ]. The GTEx GEM was preprocessed by log2 transformation of the expression values obtained from GTEx, applying the Kolmogorov–Smirnov test to remove outlier samples and quantile normalization on the GEM.…”

“…The workflow is not new, as it uses the Knowledge Independent Network Construction (KINC) software [ 47 ] and has been used in several recent plant, human and data management studies. These include identification of csGCNs that are specific to tumor types in human cancer [ 48 , 49 ], normal brain tissue [ 50 ], post-harvest-specific networks in domesticated apples ( Malus domestica ) [ 51 ], root nodulation biomarkers in Medicago [ 52 ], ripening in d’Anjou pears [ 53 ] and exploration of the effects of ‘lossy compression’ of GEMs [ 54 ]. Here, we describe sources of noise, compile a review of the workflow and describe the computational performances and limitations.…”

Brain-Region-Specific Genes Form the Major Pathways Featuring Their Basic Functional Role: Their Implication in Animal Chronic Stress Model

“…Another crucial task is identifying brain region-specific markers. A recent paper on this topic [ 8 ] utilized GTEX expression data to clarify five brain regions, specifically gene sets determined by a deep learning clustering procedure. However, there are still numerous brain sub-regions that are not included in this list.…”

“…Five brain regions employed in the study play a significant role in the chronic stress response and behavior are the Hypothalamus (HPT), Hippocampus (HPC), Dorsal Striatum (STR), Ventral Tegmental Area (VTA), and Midbrain Raphe Nuclei (MRN) [ 11 ]. These regions served as the foundation for elucidating Brain Region Specific Genes (BRSG) and their specific networks, similar to those described in [ 8 ].…”