Deep learning the dynamic regulatory sequence code of cardiac organoid differentiation.

Area of Science:

• Developmental Biology
• Genomics
• Cardiovascular Research

Background:

• Understanding human organogenesis, particularly heart development, is crucial for addressing congenital diseases.
• Temporal gene regulatory programs are key to normal embryonic development.
• Congenital heart defects often arise from disruptions in early developmental processes.

Purpose of the Study:

• To define the temporal gene regulatory programs driving human cardiac organogenesis.
• To uncover the regulatory syntax governing early heart development using advanced computational methods.
• To link developmental gene regulation to the genetic basis of congenital heart disease.

Main Methods:

• Generated a time-resolved, single-cell multi-omic atlas of human induced pluripotent stem cell-derived cardiac organoids.
• Employed deep learning models to predict chromatin accessibility from DNA sequence.
• Integrated multi-omic data (chromatin, transcriptional, genetic) to identify regulatory elements and disease variants.

Main Results:

• Discovered cell-state-specific rules of cardiogenesis and context-dependent transcription factor activities (TEAD, HAND, TBX).
• Identified distinct programs controlling cardiomyocyte lineage divergence (e.g., ventricular vs. pacemaker).
• Validated the essential role of Myocardin (MYOCD) in ventricular cardiomyocyte specification through perturbation experiments.

Conclusions:

• This study elucidates the complex gene regulatory networks underlying human heart development.
• The findings highlight early developmental origins of congenital heart disease and identify specific regulatory mechanisms.
• Provides a foundation for mechanistic understanding and therapeutic strategies for congenital heart conditions.