Authors:

Rebecca A. Henzig

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

The medullary reticular formation (mRF) of the brainstem contains neural circuits known to generate the rhythms controlling various orofacial behaviors such as breathing and swallowing [1]. Disruption of these circuits by injury, stroke, or neurodegenerative disease can disrupt these behaviors; a comprehensive understanding of these circuits is thus beneficial in addressing these conditions. However, the functional, spatial, and transcriptomic organization of the mRF is not well understood; a significant obstacle is a lack of molecular markers for distinct functional cell subpopulations within the mRF. We hypothesize that machine learning (ML) techniques applied to gene expression profile data for mRF cell populations can improve the definition of known subregions and identify new subregions involved in orofacial motor control. This research applies ML methods to the MERFISH imputed genes dataset of the Allen Brain Cell Atlas, a publicly available single-cell resolution spatial transcriptomics dataset from the Allen Institute for Brain Science that measures expression of ~8,000 genes across ~4 million cells in the whole adult mouse brain [2]. K-means clustering of 12,770 neurons mapped to the mRF by the Allen Institute's CCFv3 coordinate framework was performed using expression values for 5963 genes in the MERFISH dataset. This model generates several clusters spatially restricted to smaller subregions within the mRF, despite not including physical space as a variable. Additionally, the model isolates transcriptomically distinct motoneuron populations from other neuronal types, its clusters demonstrate compositional patterns related to neurotransmitter classes, and one cluster appears localized to the pre-Bétzinger complex (preBétC) which has been identified as the core breathing rhythm generator but has not yet been molecularly defined [3]. These results demonstrate the feasibility of this ML-oriented approach and its potential to identify candidate functional subregions of the mRF that can be further investigated using additional computational and experimental approaches.

Source:

DePaul University

Topics:

No topics listed

Co-authors:

Rebecca A. Henzig