Transcriptomic correlates of electrophysiological and morphological diversity within and across neuron types

Abstract

In order to further our understanding of how gene expression contributes to key functional properties of neurons, we combined publicly accessible gene expression, electrophysiology, and morphology measurements to identify cross-cell type correlations between these data modalities. Building on our previous work using a similar approach, we distinguished between correlations which were "class-We reason that gene-property relationships that are non-class-driven would be more likely to be potential causal regulators of the associated property. Although some class-driven correlations likely do reflect true relationships between genes and properties which distinguish excitatory from inhibitory cells, separating these relationships from instances where one cell class has a higher value of a property and coincidentally higher or lower expression of a gene without additional sources of data is not possible. Effectively, such situations are analogous to attempting to draw conclusions about correlations with only two data points.Due to limitations in available data, we were unable to address the effect of cell class in our previous work (11). Since then, the RNA-seq and electrophysiology datasets from the Allen Institute for Brain Science (AIBS) (which we originally used as validation data) have expanded greatly, with more cells and more transgenic lines represented. This increase in size, together with the fact that the AIBS data were collected using standardized protocols, suggests that this dataset might prove valuable for discovering genes correlated with electrophysiological and morphological properties. In addition, the growing use of the Patch-seq methodology (17), allowing transcriptomic, electrophysiological, and morphological characterization of the same single cell, also affords an opportunity to test gene-property correlations.Leveraging the larger size of the new AIBS dataset, we were able to address limitations of our previous study related to excitatory versus inhibitory cell class by employing statistical methods to help mitigate the effects of cell class. These methods, together with the larger number of cell types represented in the new dataset, allowed us to identify novel electrophysiological and morphological property-related gene sets which are potentially more likely to represent meaningful biological relationships.showing class-driven relationships (such as Fig 1B) but might miss some instances of non-class-driven relationships (such as Fig 1C).Another possible gene-property relationship is one where there is an interaction between gene and class, meaning that the gene-property relationship is different in excitatory and inhibitory cell types. An interaction could indicate either that excitatory and inhibitory cell types both show a correlation between the gene and property, but the slopes are in opposite directions (as in the example in Fig 1D), or that the gene is correlated with the property only in one cell class. To detect such situations, we introduced a third model, the interaction model, whic...

Cell Surface Protein mRNAs Show Differential Transcription in Pyramidal and Fast-Spiking Cells as Revealed by Single-Cell Sequencing

“…Cell-selective targeting is highly developed in cancer treatment using cell surfaceome targets (surfaceome database: http://wlab.ethz.ch/surfaceome/), but it is in infancy in the field of brain diseases. However, there are evidences about the causality between the transcriptomic and electrophysiological profile of a neuron (Cadwell et al 2016;Fuzik et al 2016;Bomkamp et al 2019) supporting the idea that development of cell typeselective fine tuning of cellular electrical activity could be promoted by single-cell transcriptomics.…”