SEE: A Method for Predicting the Dynamics of Chromatin Conformation Based on Single-Cell Gene Expression

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Chromatin conformation dynamics are crucial for cellular processes like gene regulation and DNA repair.
• Understanding these dynamics at a single-cell level is essential for deciphering complex biological functions and diseases.

Purpose of the Study:

• To introduce SEE, an AI-powered method for analyzing single-cell chromatin dynamics.
• To investigate chromatin rearrangements and interactions across different scales using SEE.
• To interpret disease-associated genetic variations by analyzing dynamic chromatin features.

Main Methods:

• Development and application of SEE, an AI method combining autoencoder and transformer techniques.
• Utilizing single-cell RNA sequencing data and limited single-cell Hi-C maps for analysis.
• Investigating chromatin dynamics at the scale of topologically associating domains (TADs) and gene loci.

Main Results:

• SEE enables high-resolution, single-cell analysis of chromatin dynamics.
• The method detected rearrangements in TADs and oscillations in chromatin interactions at gene loci.
• SEE facilitates the interpretation of disease-associated single-nucleotide polymorphisms (SNPs) through dynamic chromatin features.

Conclusions:

• SEE provides a novel, high-resolution approach to studying chromatin dynamics in single cells.
• This method offers valuable insights into both developmental processes and disease mechanisms.
• SEE enhances our ability to link genetic variations to chromatin conformational changes and cellular function.