HCMB

Last update: v1.0.1

Python code for Wu et al. HCMB: A stable and efficient algorithm for processing the normalization of highly sparse Hi-C contact data

This is an an algorithm, Hi-C Matrix Balancing (HCMB), based on iterative solution of equations, combining with linear search and projection strategy to normalize the Hi-C original interaction data.

Several useful functions:

HCMB

• main code of HCMB algorithm

get_norm_HCMB

• use HCMB to normalize Hi-C contact maps data

  input: raw contact matrix n-by-n

  output: normlaize matrix, minvalue, maxvalue, running time

NormalizeKnightRuizV2

• Python version of Knight Ruiz algorithm for matrix balancing (Knight et al. 2013). This python code is from Rajendra,K. et al. (2017)

  input: raw contact matrix n-by-n

  output: normlaize matrix, minvalue, maxvalue

and a few functions for plotting and so on.

Requirements

• Python 3
• numpy
• scipy
• matplotlib
• gcMapExplorer

Usage

from HCMB import get_norm_HCMB, NormalizeKnightRuizV2

# M is nxn matrix type: numpy.ndarray
DM, min, max, rt = get_norm_HCMB(M)
DKR, min, max = NormalizeKnightRuizV2(M)

Reference

Knight,P.A. et al. (2013) A fast algorithm for matrix balancing. IMA Journal of Numerical Analysis, 33, 1029-1047.

Rajendra,K. et al. (2017) Genome contact map explorer: a platform for the comparison, interactive visualization and analysis of genome contact maps. Nucleic Acids Research, 45(17):e152.