Resolving spatial inconsistencies in chromosome conformation measurements

Area of Science:

• Genomics
• Molecular Biology
• Biophysics

Background:

• Chromosome structure dictates function, and Chromosome Conformation Capture (3C) is key to studying it.
• 3C interaction frequencies typically correlate with spatial distances, but raw data averages across cells, introducing uncertainty.
• The spatial distances for specific interactions in 3C data remain ambiguous due to cellular heterogeneity.

Purpose of the Study:

• To develop a novel method for filtering 3C interaction data.
• To select subsets of interactions that adhere to metric constraints, improving spatial accuracy.
• To address the inherent uncertainty in 3C interaction distances.

Main Methods:

• Introduction of a new filtering technique for 3C interactions based on metric constraints.
• Application of heuristics and approximation algorithms to address computational complexity.
• Evaluation of the filtering scheme's performance against standard techniques.

Main Results:

• The metric filtering approach yields subgraphs with higher statistical significance.
• This method reduces embedding error and sensitivity to initial conditions.
• Filtered 3C data structures show improved agreement with light microscopy measurements.
• The scheme supports both strict and relaxed frequency-to-distance mappings.

Conclusions:

• The developed filtering method for 3C data integrates metric consistency and statistical confidence.
• This simultaneous consideration leads to lower-error chromosome structure embeddings.
• The resulting embeddings are more biologically plausible and reliable.