nls2_fit: =========================================================================...

In CyanolabFreiburg/rifi: 'rifi' analyses data from rifampicin time series created by microarray or RNAseq

========================================================================= nls2_fit

nls2_fit estimates decay for each probe or bin

Description

nls2_fit uses nls2 function to fit a probe or bin using intensities of the time series data from different time point. nls2 uses different starting values through expand grid and selects the best fit. Different filters could be applied prior fitting to the model.

Usage

nls2_fit(
  inp,
  cores = 1,
  decay = seq(0.01, 0.11, by = 0.02),
  delay = seq(0, 10, by = 0.1),
  k = seq(0.1, 1, 0.2),
  bg = 0.2
)

Arguments

inp	SummarizedExperiment: the input with correct format.
cores	integer: the number of assigned cores for the task.
decay	numeric vector: A sequence of starting values for the decay. Default is seq(.08, 0.11, by=.02)
delay	numeric vector: A sequence of starting values for the delay. Default is seq(0,10, by=.1)
k	numeric vector: A sequence of starting values for the synthesis rate. Default is seq(0.1,1,0.2)
bg	numeric vector: A sequence of starting values. Default is 0.2.

Details

To apply nls2_fit function, prior filtration could applied.

1. generic_filter_BG: filter probes with intensities below background using threshold. Those probes are filtered.

2. filtration_below_backg: additional functions exclusive to microarrays could be applied. Its very strict to the background (not recommended in usual case).

3. filtration_above_backg: selects probes with a very high intensity and above the background (recommended for special transcripts). Probes are flagged with "ABG". Those transcripts are usually related to a specific function in bacteria. This filter selects all probes with the same ID, the mean is applied, the last time point is selected and compared to the threshold.

The model used estimates the delay, decay, intensity of the first time point (synthesis rate/decay) and the background. The coefficients are gathered in vectors with the corresponding IDs. Absence of the fit or a very bad fit are assigned with NA. In case of probes with very high intensities and above the background, the model used makes abstinence of background coefficient. The output of all coefficients is saved in the metadata. The fits are plotted using the function_plot_fit.r through rifi_fit.

Value

the SummarizedExperiment object: with delay and decay added to the rowRanges. The full fit data is saved in the metadata as "fit_STD".

delay:

Integer, the delay value of the bin/probe

half_life:

Integer, the half-life of the bin/probe

Examples

data(preprocess_minimal)
nls2_fit(inp = preprocess_minimal, cores = 2)

CyanolabFreiburg/rifi documentation built on May 7, 2023, 7:53 p.m.