AUTHOR=Keller Daniel , Meystre Julie , Veettil Rahul V. , Burri Olivier , Guiet Romain , Schürmann Felix , Markram Henry 

TITLE=A Derived Positional Mapping of Inhibitory Subtypes in the Somatosensory Cortex

JOURNAL=Frontiers in Neuroanatomy

VOLUME=Volume 13 - 2019

YEAR=2019

URL=https://www.frontiersin.org/journals/neuroanatomy/articles/10.3389/fnana.2019.00078

DOI=10.3389/fnana.2019.00078

ISSN=1662-5129

ABSTRACT=Obtaining a catalog of cell-types is a fundamental building block for understanding the brain. The ideal classification of cell-types is based on the profile of molecules expressed by a cell, in particular the profile of genes expressed. One strategy is therefore to obtain as many single-cell transcriptomes as possible and isolate clusters of neurons with similar gene expression profiles. In this study, we explored an alternative strategy. We explored whether cell-types can be algorithmically derived by combining protein tissue stains with transcript expression profiles on 3D space. We developed an algorithm that aims to distribute cell-types in the different layers of a column of cortical tissue of the developing rat constrained by the tissue- and cellular level data. We found that the spatial distribution of major inhibitory cell types can be approximated using the available data. The result is a mini atlas of inhibitory cell-types of the rat somatosensory cortex. In principle, any data that constrains what can occur in a particular part of the brain can also strongly constrain the derivation of cell-types. This draft inhibitory cell-type mapping is therefore dynamic and can iteratively converge towards the ground truth as further data is integrated. It is made available as a public resource and open for collaborative improvement. The approach developed should scale to the whole brain. 
Neurons are known to express particular combinations of molecules, which can be used as markers to classify different cell types (Rudy et al. 2011). However, we still lack actual numbers of each type of interneuron through the cortical depth. This is largely due to the time-consuming methods classically used, which rely on large-scale sampling and characterization of cellular composition. To overcome these challenges and to systematically describe cellular composition throughout the cortical depth, we apply a marker-fitting method that combines immunohistochemistry data with RNA transcript expression (reverse transcription polymerase chain reaction, RT-PCR) profiles of individual cells to predict interneuron cell types. We show the layer-dependent distribution of major inhibitory cell types can be predicted in P14 rat somatosensory cortex, based on protein and transcriptome markers.