HiCForecast: dynamic network optical flow estimation algorithm for spatiotemporal Hi-C data forecasting.

Area of Science:

• Genomics
• Computational Biology
• Molecular Biology

Background:

• Spatiotemporal Hi-C data reveals DNA's 3D organization during cellular development.
• Limited data availability hinders the study of 3D genome dynamics across species.
• Predictive methods are needed to generate future Hi-C data from existing time-series datasets.

Purpose of the Study:

• To develop a novel computational framework for forecasting spatiotemporal Hi-C data.
• To evaluate the generalizability and performance of the forecasting method across diverse biological systems.
• To provide a tool that enhances the study of 3D genome organization dynamics.

Main Methods:

• Developed HiCForecast, a framework utilizing a dynamic voxel flow algorithm.
• Assessed forecasting performance across homogeneous, heterogeneous, and general contexts.
• Employed computational and biological metrics for rigorous evaluation.

Main Results:

• HiCForecast demonstrates superior performance compared to the current state-of-the-art.
• The method shows effective generalization across different species and systems.
• HiCForecast provides accurate predictions of future spatiotemporal Hi-C datasets.

Conclusions:

• HiCForecast is an efficient and powerful tool for generating future spatiotemporal Hi-C data.
• The framework addresses the limitation of sparse Hi-C data, advancing 3D genome organization studies.
• The tool is publicly available to facilitate further research.