Mapping the Molecular Landscape of ALS Through Integrative Omics and Network Pharmacology

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder characterised by progressive motor neuron loss and marked molecular heterogeneity. This thesis presents a two-part computational multi-omic framework utilising the Answer ALS cohort and publicly available single-cell RNA sequencing data to identify early-stage biomarkers and nominate cell-type-specific therapeutic candidates through network guided drug repurposing. In Part I, Weighted Gene Co-expression Network Analysis (WGCNA) of bulk RNA- seq and ATAC-seq data from the Answer ALS cohort revealed that biological sex is the dominant source of transcriptomic variance, completely masking disease-associated signals in unsupervised global analyses. Sex-stratified network construction identified a male-specific co-expression module enriched for non-coding RNAs at the DLK1–DIO3 locus, from which LASSO logistic regression distilled a seven-gene diagnostic signa- ture with a cross-validated AUC of 0.719. Consensus clustering of chromatin accessi- bility data identified a spurious two-cluster structure driven by Y-chromosomal mapping artifacts, which resolved upon peak removal. Differential correlation analysis of paired ATAC-seq and RNA-seq data further revealed a significant regulatory inversion at the EIF5 locus, implicating dysregulated RAN translation as a potential upstream amplifier of C9orf72-associated pathology. In Part II, publicly available single-cell RNA sequencing data from TDP-43 ALS cortex was used to construct cell-type-specific gene regulatory networks. Network proximity-based drug repurposing identified significant candidates in the microglial and oligodendrocyte modules, while the astrocyte and excitatory neuron networks lacked sufficient topological density to nominate candidates. In microglia, the top-ranked com- pounds converged on GABA-A receptor subunits, with additional pharmacologically distinct candidates including venlafaxine, nortriptyline, pregabalin, and iloperidone. In oligodendrocytes, the dominant signal was dopamine D2 receptor antagonism, with ropinirole, risperidone, ziprasidone, and progesterone representing the highest-priority candidates. Together, these findings demonstrate that rigorous confounder correction, sex- stratified analyses, and cell-type-specific network frameworks are essential for extract- ing biologically meaningful signals from population-scale ALS omics data.