Characterising Early Neurodevelopmental Effects and Post-Transcriptional Dysregulation in FUS and TARDBP iPSC Models of ALS
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder lacking an effective cure. While mutations in RNA-binding proteins (RBPs) like FUS and TDP-43 (encoded by TARDBP) are known to cause pathogenic nucleocytoplasmic mislocalisation and downstream motor neuron degeneration, the early developmental molecular deviations remain largely uncharacterised. This project aimed to uncover these early differences in transcriptomic signatures by analysing a single-cell RNA-sequencing (Smart-Seq2) dataset of developing human iPSC-derived neurons.
Using a computational workflow comprising differential gene expression, alternative splicing quantification, and pseudotime trajectory inference, wild-type cells were compared against FUS and TARDBP mutants at two biological resolutions. Compositional analysis indicated significant proportional shifts in mutant progenitors and motor neurons, while interneurons remained largely unaffected. Static differential expression profiling demonstrated that mutations target cell-specific pathways rather than housekeeping functions, with motor neurons primarily showing a transcriptomic upregulation. Pairwise splicing analysis revealed unique cell-type-specific disruptions alongside a core cohort of 49 convergent splicing events when mutations were grouped together. Finally, trajectory modelling revealed lineage-specific dynamics. Across top dynamic genes, interneuron pathways showed uniform transcriptomic suppression, whereas mutant motor neurons exhibited distinct developmental patterns, including severe activation delays, uniform downregulation, and sustained or early overexpression. Ultimately, all results point to a clear change in the developmental time window of the iPSC models, indicating that, despite being a late-onset disease, the disease's foundational disruptions begin during development.
/Originals: Untouched raw data (Smart-seq2)./Metadata: SraRunTable and extracted SRR IDs./Scripts: Executable code (Numbered00_to14_in execution order)./Results:/Outputs: Outputs generated from analysis (Whippet, Trade-seq, PHATE, STAR alignments)./Figures: Final plots generated during analysis (UpSet plots, Treemaps, etc.).
The following software, tools, and R packages are required to replicate the computational workflow:
- SRArun tool (v3.0.3-gcc-13.2.0)
- FastQC (v0.12.1-gcc-13.2.0)
- STAR aligner (v2.7.10b)
- MultiQC (v1.15)
- R (v4.3.0 / 2023-04-21) and RStudio "2024.4.999.999"
- Seurat (v5.0.1)
- rtracklayer (v1.64.0)
- SingleCellExperiment (v1.24.0)
- scDblFinder (v1.16.0)
- clustree (v0.5.1)
- Ducasse (v0.99.0)
- UpSetR (v1.4.0)
- gprofiler2 (v0.2.3)
- slingshot (v2.10.0)
- tradeSeq (v1.16.0)
The analysis is broken down into 14 sequential scripts located in the /Scripts directory. They must be executed in the following order to successfully process the raw data through to the final trajectory models:
The full transcript coverage (SmartSeq2) single-cell sequencing data used in this study originates from the publication "Single-cell RNA-sequencing reveals early mitochondrial dysfunction unique to motor neurons shared across FUS- and TARDBP-ALS" by C. Schweingruber et al. (2025).
The raw data and the sample metadata (SraRunTable.csv) can be accessed and downloaded from the NCBI SRA Run Selector using the Project ID: PRJNA940433 or GSE226482.