filter_params: Determine the A quantile cutoff to be used

In dozmorovlab/HiCcompare: HiCcompare: Joint normalization and comparative analysis of multiple Hi-C datasets

Determine the A quantile cutoff to be used

Description

Determine the A quantile cutoff to be used

Usage

filter_params(
  hic.table,
  SD = 2,
  numChanges = 300,
  FC = 3,
  alpha = 0.05,
  Plot = FALSE
)

Arguments

hic.table	A hic.table object
SD	The standard deviation of the fuzzing used to produce a Hi-C matrix from your data with few true differences.
numChanges	The number of changes to add into the Hi-C matrix created. This should be proportional to the resolution of the data. High resolution data should use more changes i.e. 1MB resolution - 300 changes, 100KB resolution - 1000 changes, etc.
FC	The fold change of the changes added to the Hi-C matrix.
alpha	The alpha level for hypothesis testing.
Plot	logical, should MD plots for the normalization and difference detection be plotted?

Details

This function will take your data and produce an additional Hi-C matrix using the IF1 vector. Random normal noise will be added to the vector to create a "fuzzed" matrix with few true differences. Then the specified number of true changes will be added at the specified fold change level to the matrices. The HiCcompare procedure is run on the data and a plot of the MCC, TPR, and FPR based on the A minimum value filtered out will be produced. This is to aid you in determining what value you should use when analyzing your data with the hic_compare() function.

Value

A plot of the Mathews Correlation Coefficient (MCC), true positive rate (TPR), and false positive rate (FPR) over the A minimum value filtered.

Examples

data('HMEC.chr22')
data('NHEK.chr22')
hic.table <- create.hic.table(HMEC.chr22, NHEK.chr22, chr = 'chr22')
filter_params(hic.table)

dozmorovlab/HiCcompare documentation built on June 30, 2023, 3:09 a.m.