Exploratory Analyses
Differential Contact Maps Based on Gene Expression Changes
Is there a tool that can generate a differential chromatin contact map based on expression changes, given that a control condition and its corresponding contact matrix are available?
I’m exploring whether any existing tool can integrate gene expression changes with chromatin contact data to generate condition-specific or differential contact maps
CLOD Reimplementation
todo - add the paper here
To identify long-range chromatin contacts exhibiting condition-specific enrichment, I reimplemented the CLOD (Chromosome Long-range colocalization identifier through Outlier Detection) method from Zhang et al., 2024.
The method operates as follows:
- Smooths the Hi-C contact matrix using a uniform filter
- Computes genomic distances between all contact pairs
- Groups contacts into bins based on separation distance
- Within each distance bin:
- Computes the third quartile (Q3) and interquartile range (IQR)
- Flags contacts as outliers if:
Contact signal > Q3 + 2 × IQR
This outlier-based detection is designed to isolate statistically enriched, high-contact interactions.
I applied this pipeline at 10 kb resolution on the balanced Hi-C matrix of chromosome 14.
This minimal yet functional reimplementation captures the statistical foundation of CLOD
Evaluation of CLOD-Detected Contacts
While CLOD detects statistical outliers, the corresponding Aggregate Peak Analysis (APA) plot shows low to modest central enrichment.
- APA reflects the mean behavior of the contact set
- Signal heterogeneity across contacts(esp. long distance) may blunt the APA signal
- Variable interaction patterns can reduce sharpness despite real biological significance
Additional Analyses
Motif Enrichment
(todo)
GC Content
(todo)