compute_histogram: Compute a quality metric.

In angelesarzalluz/scfilters: A correlation-based method for quality filtering of single-cell RNAseq data

Description

Computes a statistical metric to assess differences between data distribution of correlation density and negative control.

Usage

compute_histogram(density_data)

Arguments

density_data

A data frame containing the density data.

Details

The aim of this function is to compute the ratio of correlation density that is non- coincident with the density distribution of the negative control. Both distributions are tested to find the section where data density distribution is greater than the control's, and the differences between both are computed. This metric is divided by the total data density of correlation.

Graphically, this is the computation of the area where the correlation density distribution of the data does not overlap with that of the negative control, divided by the total area under the data correlation density curve. The aim is to determine which windows present a significant similarity to a random distribution, and therefore are highly affected by systematic noise, in order to discard them.

The metric is computed by the mean control curve, as well as for the maximum and minimum curves, obtaining the error values. The returned data frame containis the following three columns:

• hist_value: the result of computing the metric for the mean control curve.

• error_up: the value of the metric for the maximum control curve.

• error_down: the value of the metric for the minimum control curve.

Important note: the compute_histogram function was designed to operate on the output of the compute_density function.

Value

A data frame containing the value of the metric for each window, and values for error bars to plot them as a histogram.

angelesarzalluz/scfilters documentation built on May 10, 2019, 11:46 a.m.