aggregate_diagonals: Aggregates diagonals of based on similarity of XY...

In rz6/DIADEM: Hi-C differential analysis via dependency modelling of chromatin interactions with generalized linear models

Description

Merges 2 contact maps in sparse format and then aggregates it diagonals. The aggregation process is performed as follows:

• start with k = 1 (first diagonal), group k consists of observations: val.x * val.y (from k-th diagonal)

• take merge candidate i.e.: k' = k + 1, , group k' consists of observations: val.x * val.y (from k'-th diagonal)

• perform chi-square test between group k and k', H0: observations from both groups come from the same distribution, H1: they come from different distributions

• if rejected set k = k' (and pool diagonals k to k'-1 into single group), otherwise set k' = k' + 1

• repeat until all diagonals are processed

After algorithm is finished diagonals are assigned to groups (pools).

Usage

aggregate_diagonals(
  mtx1.sparse,
  mtx2.sparse,
  agg.diags = TRUE,
  which.test = c("energy", "KS")[1],
  alpha = 0.05,
  exclude.outliers = FALSE
)

Arguments

mtx1.sparse	data frame, contact map in sparse format
mtx2.sparse	data frame, contact map in sparse format
agg.diags	logical if false leave each diagonal in separate group (no pooling)
alpha	numeric significance threshold for chi square test

rz6/DIADEM documentation built on Dec. 31, 2019, 3:51 a.m.