Fig. 3: Integrative analysis of multiple transcriptomics datasets reveal Parkinson’s disease-related gene signatures and biological processes in patient-reprogrammed neurons.
From: Druggable transcriptomic pathways revealed in Parkinson’s patient-derived midbrain neurons
A Heatmap of mean log2 fold changes in expression of the top 20 genes with the highest dysregulation across all six datasets of PD versus control reprogrammed neurons. Red and blue indicate, respectively, up- and downregulation in PD cells relative to control. Gene function and overall dysregulation score (Doverall) are annotated vertically according to the legend. Dsc-iPS, Dbulk-iPS and Dbulk-iN scores indicate the strength of dysregulation for, respectively, the two single-cell iPSC neuron datasets (orange), the two bulk iPSC neuron datasets (blue) and the two bulk iN datasets (green) (see Methods for details). *, nominal Pcombined < 0.05. B, C, D ToppFun functional enrichment analysis of the 200 most highly dysregulated genes across all datasets highlights fundamental biological processes (B), cellular components (C) and chemicals (D) implicated or suspected to be involved in PD. Chord plot in (B) indicates genes annotated to dysregulated GO biological processes, ordered by decreasing overall dysregulation score. Length of bars in (C, D) represent Benjamini-Hochberg-corrected significance values and numbers indicate number of genes annotated to GO term; grey color indicates significance threshold. Extended data in Supplementary Tables 4 and 5. E Gene Set Enrichment Analysis (GSEA) of multiple-dataset PD dysregulated genes for genes associated with the terms “Parkinsonism”, “Dementia” and “Schizophrenia” from the Human Phenotype Ontology (HPO) database131. “Parkinsonism”- and “Dementia”-related genes were significantly overrepresented at the top of the list of expressed genes (n = 24,693) rank-ordered by decreasing Doverall score. F GSEA shows a significant enrichment of PD-symptom-related HPO terms “Bradykinesia”, “Rigidity”, “Akinesia”, “Dyskinesia” and “Postural instability” among the multiple-dataset PD dysregulated genes. Vertical bars in (E, F) represent the “gene hits”, i.e., the location of genes from each indicated HPO term within the Doverall rank-ordered list. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant (P > 0.05); all P-values are FWER-corrected to exclude any possibility of false-positive enrichment. NES Normalized Enrichment Score. Detailed enrichment results are provided in Supplementary Table 6.
