Normalization of Hi-C maps

To normalize the Hi-C maps, several methods are implemented.

  • Iterative Correction (IC) [1] This method normalize the raw contact map by removing biases from experimental procedure. This is an method of matrix balancing, however, in the normalized, sum of rows and columns are not equal to one.
  • Knight-Ruiz Matrix Balancing (KR) [2] The Knight-Ruiz (KR) matrix balancing is a fast algorithm to normalize a symmetric matrix. A doubly stochastic matrix is obtained after this normalization. In this matrix, sum of rows and columns are equal to one.
  • Vanilla-Coverage (VC) [3] This method was first used for inter-chromosomal map. Later it was used for intra-chromosomal map by Rao et al., 2014. This is a simple method where at first each element is divided by sum of respective row and subsequently divided by sum of respective column.
  • Median Contact Frequency Scaling (MCFS) This method can be used to normalize contact map using Median contact values for particular distance between two locations/coordinates. At first, Median distance contact frequency for each distance is calculated. Subsequently, the observed contact frequency is divided by median contact frequency obtained for distance between the two locations.

To perform these normalizations, following tools are implemented:

References

[1]Imakaev et al. Iterative correction of Hi-C data reveals hallmarks of chromosome organization. Nature Methods 9, 999–1003 (2012).
[2]Knight P and D. Ruiz. A fast algorithm for matrix balancing. IMA J Numer Anal (2013) 33 (3): 1029-1047.
[3]Lieberman-Aiden et al. Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science (2009) 326 : 289-293.