Background & Aims

Peripheral primary sensory neurons detect, filter, transduce and transmit relevant external and internal sensory information to the CNS. These specialized neurons have their cell bodies located within primary sensory ganglia, where various non-neuronal cell types provide essential support and nurturing. Given their significance in therapeutic applications, recent efforts have been directed towards performing single-cell transcriptomic analyses on these ganglia. In addition to quantitative differences at the level of mRNA counts, new long-read sequencing technologies allow the identification of specific isoforms in various cell types. In this study we use improved integration techniques and novel statistical models for inferring biological differences between donor populations. We also aim to use these statistical models to unravel differential enrichment of isoforms in every cell type present in human dorsal root ganglia (DRGs).

Methods

The increasing accumulation of sparse “omics” datasets necessitates the development of novel statistical models capable of effectively mitigating experimental variabilities and “batch” effects. To systematically validate and enhance the integration and interpretation of multimodal data, we have devised a hierarchical study design for the collection and analysis of single-nucleus transcriptomic data from postmortem human DRGs. Samples from multiple human donors are each subjected to different dissociation protocols and various capture chemistries. Sequencing results are meticulously preprocessed and integrated through the utilization of a Bayesian generative model. Implementation of a nested linear model enables the estimation of both random effects (such as technical variabilities) and fixed effects (such as the sex and age of the donors) in a cluster-free analysis of differential gene and isoform expression.

Results

Lumbar DRGs from 7 healthy donors, 3 male and 4 females, were dissociated with 2 different protocols and 3 different chemistries (total 36 captures). After sequencing quality control, background removal, and doublet calling, 179041 nuclei passed the QC requirements and were included for further analysis. The data from these nuclei were integrated and compared to previously published datasets [Jung et al., 2023] bringing the total dataset close to 3.105 cells. Five different integration methods were compared, and cell types were defined based on the highest integration method score. Differential gene expression analysis reveals the presence of ~15 main cell types including neurons, satellite glial cells, macrophages, and blood cells. Long-read sequencing of the single cell libraries (39000 cells, 2000 median read per cell) reveals the presence of differentially expressed isoforms that require further validation through other experimental modalities.

Conclusions

Single cell “omics” data are essential for screening, target identification, and validation of tissue-specific therapeutic targets. In our study we aim to generate a comprehensive atlas that allows researchers in the field of somatosensation, more specifically pain, to validate new strategies for targeting specific cell types and/or for selecting key pathways that contribute to sensory perception. Our integration and analysis pipelines pave the way for the addition of other datasets and modalities. We hope this atlas will be a powerful tool to be used by many at various stages of research and development.

References

Jung, M., Dourado, M., Maksymetz, J., Jacobson, A., Laufer, B.I., Baca, M., Foreman, O., Hackos, D.H., Riol-Blanco, L., and Kaminker, J.S. (2023). Cross-species transcriptomic atlas of dorsal root ganglia reveals species-specific programs for sensory function. Nat. Commun. 14, 366. 10.1038/s41467-023-36014-0.

Presenting Author

Behrang Sharif

Poster Authors

Behrang Sharif

PhD

McGill University

Lead Author

E-mail

Spyros Oikonomopoulos

McGill Genome Centre, Department of Human Genetics, McGill University

Lead Author

E-mail

Yu Wang

McGill Genome Centre, Department of Human Genetics, McGill University

Lead Author

E-mail

Haig Djambazian

McGill Genome Centre, Department of Human Genetics, McGill University

Lead Author

E-mail

Eve Tsai

Division of Neurosurgery, University of Ottawa

Lead Author

E-mail

Michael Hildebrand

Department of Neuroscience Carleton University

Lead Author

E-mail

Jiannis Ragoussis

McGill Genome Centre, Department of Human Genetics and Bioengineering, McGill University

Lead Author

E-mail

Philippe Séguéla

Montreal Neurological Institute, Department of Neurology & Neurosurgery, McGill University

Lead Author

E-mail

Topics

  • Genetics