scHiCNorm

scHiCNorm applies zero-inflated and hurdle statistical models to remove systematic biases from single-cell Hi-C datasets, enabling more accurate detection of cell-to-cell variation in chromosomal structures.

Key Features:

• Statistical Modeling: Utilizes zero-inflated Poisson (ZIP), zero-inflated Negative Binomial (ZINB), hurdle Poisson (HP), and hurdle Negative Binomial (HNB) models tailored for zero-inflated single-cell Hi-C contact counts.
• Bias Removal: Corrects systematic biases arising from cutting sites, GC content, and mappability in Hi-C contact data.
• Scripted Feature Generation and Correction: Provides Perl scripts to generate bias features and R scripts to apply the statistical models for bias correction.
• Pre-built Reference Bias Features: Includes pre-built bias feature files for human (hg19, hg38) and mouse (mm9, mm10) reference genomes.

Scientific Applications:

• Single-cell chromatin conformation analysis: Enables more reliable interpretation of chromosomal contact maps from single-cell Hi-C experiments.
• Cellular heterogeneity studies: Facilitates detection of genuine cell-to-cell differences in chromosomal architecture.
• Development and disease research: Supports studies of development, disease progression, and therapeutic responses by reducing technical artifacts in single-cell Hi-C data.

Methodology:

Generate bias features using Perl scripts and apply statistical models (ZIP, ZINB, HP, HNB) via R scripts to correct contact counts; includes pre-built bias features for hg19, hg38, mm9, and mm10.