Neocortical Development

Drawing on BRAIN Initiative and other public resources, we have assembled hundreds of highly curated studies, including millions of samples, into a collection of ready-to-explore transcriptomic data focused on neocortical development (bioRxiv). The compendium spans in vivo development across mammalian species, as well as in vitro stem cell models of cortical development. Access through our biologist-friendly NeMO Analytics data exploration environment is entirely open and aims to give researchers the ability to visualize and explore many experiments simultaneously without coding expertise: Neocortical Development at NeMO Analytics

To harness the combined discovery potential of the compendium, we have applied our structured joint matrix decomposition algorithms (SJD) to focused sub-collections of data. We define conserved transcriptomic elements of neocortical neurogenesis as well as expression modules that are unique to the human lineage. The animation at left shows conserved seqential transcriptomic dynamics across mouse, macaque and human neocortical neurogenesis that were defined in this way. Leveraging shared signals across many datasets we uncover 1] a temporal disconnection between the expression of neuronal identity-defining transcription factors and the layer-specific transcriptomic programs they give rise to (linked to Nano et al preprint), 2] an apparent exception to the inside-out order in the development of neocortical layers (linked to Huilgol et al preprint), 3] insights into the evolution of specific gene regulatory networks underlying the emergence of outer, or basal, radial glial cells, and 4] which layer-specific neuronal maturation signatures are and are not robustly recapitulated in cerebral organoid systems.

By implementing transfer learning methods within NeMO Analytics, we enable researchers to incorporate these shared dimensions of transcriptome variation that we have defined into their own exploration of neocortical development. It is also possible for researchers to integrate their own emerging datasets and gene signatures of interest into this environment. We hope this unique combination of ready-to explore data and analysis tools in a graphical user interface will empower researchers to design and execute their own hypothesis-driven in silico experimentation. It is our aim that NeMO Analytics become an open data and communication hub for the neocortical development research community.

Shreyash Sonthalia lead the analysis and the NeMO Analytics resource construction for this projcet, and has compiled a comprehensive list of neocortical datasets assembled in NeMO Analytics.

We have a number of projects exploring neocortical development:

A collaboration with Gabriel Santpere, Nenad Sestan, Pasko Rakic, Flora Vaccarino to explore neurodevelopmental risk gene expression in human neural stem cells, with 1st author Xoel Mato-Blanco and senior author Nicola Micali.

CellCover: a new approach to define marker gene panels that distingush cell types in scRNA-seq data, which we use here to explore cell types in the developing neocortex and temporal progression & evolution of neural progenitor cells. This is a collaboration with Don Geman and Laurent Younes at JHU Applied Math, lead by Lanlan Ji and An Wang.

Heteregeneity of neuronal identity in NGN2-induced neurons: A collaboration with Dimitri Avramopoulos.

2011 Nature paper: Bulk gene expression analysis in the DLPFC across the human lifespan