CoGTEx

CoGTEx provides access to system-level and tissue-level coexpression estimations derived from GTEx human data, enabling exploration and download of coexpressed genes for analyses of biological pathways and identification of novel functional gene partners.

Key Features:

• Dual-level coexpression estimation: CoGTEx computes coexpression at both system-level (TPM scale, without tissue standardization) and tissue-level (z-score scale, with tissue standardization).
• Robust correlation metrics: Three correlation metrics are calculated: Pearson, Spearman, and G-statistic.
• Curated tissue clusters: PCA, clustering, and tSNE stringent procedures were applied to generate 42 distinct and curated tissue clusters from GTEx v8 expression data.
• Resampling approach: Coexpression is estimated by sampling 70 samples per tissue cluster, repeated 20 times, with the minimum value extracted as a robust estimation to avoid tissue overrepresentation.
• Gene coverage: Correlation is computed for 33,445 genes across the 42 tissue clusters.

Scientific Applications:

• Biological pathway analysis: CoGTEx facilitates analyses of biological pathways through coexpression estimation.
• Novel functional gene partner identification: The resource enables identification of novel functional gene partners via coexpression patterns.
• Transcriptional regulation analysis: Coexpression estimations are associated with transcriptional regulation.
• System-level versus tissue-level comparison: Users can compare system-level and tissue-level coexpression estimations, which contain distinct and complementary valuable biological information.

Methodology:

GTEx v8 expression data was globally normalized, batch-processed, and filtered. PCA, clustering, and tSNE stringent procedures were applied to generate 42 distinct and curated tissue clusters. Coexpression was estimated from these clusters by computing correlation of 33,445 genes, sampling 70 samples per tissue cluster repeated 20 times, extracting the minimum value as a robust estimation. Three metrics (Pearson, Spearman, and G-statistic) were calculated in two modes: system-level (TPM scale) and tissue-level (z-score scale).