Mari Sepp, Associate Professor, University of Tartu

Mari Sepp, Associate Professor, University of Tartu

Title history
  • Mari Sepp, Associate Professor, University of Tartu — observed on 20261028-0747 UTC (YouTube)

by Mari Sepp · View the channel on PSI

Life & Natural Systems Science — General & History GTInterdisciplinary studiesGTZGeneral studies and General knowledge
🎙 Mari Sepp 👥 1K 📅 October 28, 2026 ⏱ 18 min 👁 8 📄 original study 🧭 2026-10-28
Available in: English (current) Français

Keywords

cerebellumevolutionary genomicssingle-cell RNA sequencingATAC-seqautism spectrum disorders

Summary

Mari Sepp presents research on the cellular and gene regulatory programs underlying cerebellum development across mammalian evolution, with implications for neurodevelopmental disorders. The talk begins with an introduction to evolutionary genomics and the question of how diverse phenotypes arise from genotypes. She highlights that the brain evolves more slowly than other organs, yet the cerebellum shows remarkable morphological variation across species. Sepp then focuses on the cerebellum, which contains 80% of all neurons in humans, and its roles in motor control, cognition, and social behavior. The core of the talk utilizes single-nucleus RNA sequencing data from human, mouse, and opossum to create a consensus cell type classification of the developing cerebellum. Key findings include differences in Purkinje cell generation dynamics, with humans having more early-born Purkinje cells, and numerous examples of human-specific gene expression gains, such as DSCAM in specific neuronal subtypes. To investigate regulatory mechanisms, she uses single-nucleus ATAC-seq across six species and deep learning models to predict chromatin accessibility from sequence. These models were applied to 200 eutherian genomes, revealing regulatory elements that emerged at different evolutionary timepoints. One example is a primate-specific enhancer validated experimentally. The talk concludes with future directions on how human-specific aspects of cerebellum development may contribute to autism spectrum disorder.

207 words

Critical Evaluation

Value of the Information & Strength of the Argument

The information presented is highly valuable for understanding the evolutionary dynamics of the cerebellum and the molecular basis of human-specific brain features. The argumentation is solid, grounded in large-scale comparative genomics and functional validation experiments. The speaker clearly explains the logic from hypothesis to data generation and interpretation, and addresses potential caveats such as the conservation of the cerebellum. The inclusion of deep learning models to extrapolate across many species demonstrates an innovative approach. However, the talk is technical and assumes a background in genomics, limiting its accessibility to a broader audience.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is high, with detailed methods and validation steps. The speaker acknowledges collaborations (e.g., Stein Arts) and uses publicly available data to support conclusions. No explicit external sources are cited in the video or description, but the research is likely published in peer-reviewed venues. The title is only the speaker’s name and affiliation, which is common for seminar talks, and does not mislead about the content. Absence of detailed citations within the talk is typical for conference presentations.

187 words

Title / Content Match

The title accurately identifies the speaker and venue, but does not reflect the specific topic of the talk. It is a standard format for conference presentations.

Quality & Reliability

8/10

The talk presents original research from the speaker's lab and collaborators, with detailed methodology and data analysis. The speaker is an established researcher. The content is specific and technically sound, though external sources are not cited within the video itself.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Contribution & Novelties

The talk presents novel data from a cross-species single-nucleus atlas of the developing cerebellum, revealing human-specific gene expression and regulatory elements. The use of deep learning to extrapolate across 200 mammalian genomes is innovative and provides a framework for studying evolution. Future work on autism spectrum disorder leverages these evolutionary findings.

Pour aller plus loin :

  • Single-cell RNA sequencing — The core technology used to profile cell types.
  • ATAC-seq — Method to map chromatin accessibility, crucial for identifying regulatory elements.
  • DSCAM gene — A gene implicated in Down syndrome and autism, highlighted as a human-specific expression gain.
  • Autism spectrum disorder — The clinical context for ongoing research.

107 words

Radar Profile

The radar profile shows high technicality and reliability, but moderate quantity of information and somewhat lower accessibility due to the specialized topic. The speaker excels in presenting original data with strong methodology.

Reliability 8/10