ShRec3D-Exented

Shortest Path Reconstruction algorithm extended with different dimensionality reduction methods and tunable imputation parameter [1].

ShRec3D

Algorithm developped by Lesnes et al [2] which reconstruct the 3D coordinates of a chromosom from a normalized HiC map (SCN [3]). The Algorithm is basically in two parts:

• Conversion from normalized HiC contact matrix in distance matrix such that
• Conversion of the distance mtrix in 3D coordinates with MDS

Improvements

Tuning the imputation parameter

Conversion from normalized HiC contact matrix in distance matrix such that . The modification of the parameter allows to tune the range of contacts to imputes (see Video Evolution imputation (in black) with alpha )

Changing the dimensionality reduction algorithm

The MDS allows to reconstruct large distances efficiently, whereas sometimes the short range distances reconstruction are needed. The Sammon mapping and other DR (Dimensionality Reduction) algorithms allows to change the reconstruction focus.

[1] Morlot, Jean-Baptiste, Julien Mozziconacci, and Annick Lesne. "Network concepts for analyzing 3D genome structure from chromosomal contact maps." EPJ Nonlinear Biomedical Physics 4.1 (2016): 2.
[2] Lesne, Annick, et al. "3D genome reconstruction from chromosomal contacts." Nature methods 11.11 (2014): 1141-1143.
[3] Cournac, Axel, et al. "Normalization of a chromosomal contact map." BMC genomics 13.1 (2012): 436.\