Orateur
Description
Over the past decade, imaging-based chromatin tracing has emerged as a reliable method to simultaneously visualise the spatial organisation of chromatin and transcription of specific genomic loci at single-cell resolution. However, extracting interpretable structural patterns from these datasets remains a major challenge. To address this, our lab recently developed 3DTopic, a topic modelling framework that identifies recurrent chromatin folding motifs (CFMs) across single alleles in an unsupervised manner (Messina et al., 2024, bioRxiv preprint). Here, we introduce the associated software package, which provides a well-documented and accessible pipeline for the analysis of chromatin tracing data. We then demonstrate the performance of the method on (1) simple in silico datasets generated using polymer simulations, and (2) experimental data from Bintu & Chen et al., 2018, where it robustly recovers known structural features. Finally, we highlight the role of loop extrusion in the formation of such features by comparing data from wild-type and RAD21-depleted cells, uncovering distinct shifts in the folding motifs known to be associated with cohesin activity.
References:
Messina, O., Fiche, J. B., Youhanna, C. E., Ganesh, G., Kemkem, Y., Mesnager, A., ... & Nollmann, M. (2024). Chromatin folding motifs define the basic grammar of TAD architecture in single alleles. bioRxiv, 2024-09.
Bintu, B., Mateo, L. J., Su, J. H., Sinnott-Armstrong, N. A., Parker, M., Kinrot, S., ... & Zhuang, X. (2018). Super-resolution chromatin tracing reveals domains and cooperative interactions in single cells. Science, 362(6413), eaau1783.