/SELFISH

Selfish: Discovery of Differential Chromatin Interactions via a Self-Similarity Measure

Primary LanguageMATLAB

Selfish

SELFISH (Discovery of Differential Chromatin Interactions via a Self-Similarity Measure) is a tool for finding differential chromatin interactions between two Hi-C contact maps. It uses self-similarity to model interactions in a robust way. For more information read the full paper: Selfish: Discovery of Differential Chromatin Interactions via a Self-Similarity Measure. A Python implementation of Selfish is available at Ay-lab.

Dependency

  • MATLAB 8.5 or newer.

Syntax

  • [X,Y,P] = SELFISH_DCI(contc1,contc2,norm1,norm2,THRESHOLD,RESOLUTION,INTERVAL) returns the coordinates of differential chromatin interactions [X,Y] and their correponding corrected p-values (after fdr) in P for raw contact maps read from files contc1 and contc2 with normalization vectors read from files norm1 and norm2.

  • [X,Y,P] = SELFISH_DCI(contc1,contc2,[],[],THRESHOLD,RESOLUTION,INTERVAL) returns the coordinates of differential chromatin interactions [X,Y] and their correponding corrected p-values (after fdr) in P for contact maps read from files contc1 and contc2.

Parameters

SELFISH_DCI doesn't take any parameters as input but it has two predefined parameters work best for 5kb resoultion.

  • sigma0 = 1.6 first Gaussian filter sigma (r0 = 2ceil(2sigma0)+1).

  • s = 10 the number of Gaussian radii used in the modeling.

Inputs

  • contc1 Hi-C contact map filename 1 (e.g. '/chr1/MAPQGE30/chr1_5kb.RAWobserved')

  • norm1 normalization vector filename 1 (e.g. '/chr1/MAPQGE30/chr1_5kb.KRnorm')

  • contc2 Hi-C contact map filename 2

  • norm2 normalization vector filename 2

  • THRESHOLD fdr threshold used for calling DCIs (e.g. 10^-4)

  • RESOLUTION data resoultion in bp (e.g. 5000)

  • INTERVAL the interval in bp for which DCIs are detected (e.g. [10000000 35000000])