model_bam_standards: Build a Bayesian additive model from spike-ins to correct...

In trichelab/spiky: Spike-in calibration for cell-free MeDIP

Build a Bayesian additive model from spike-ins to correct bias in *-seq

Description

Build a Bayesian additive model from spike-ins to correct bias in *-seq

Usage

model_bam_standards(x, conc = NULL, fm = NULL, ...)

Arguments

x	data with assorted feature information (GCfrac, CpGs, etc)
conc	concentration for each spike (must be provided!)
fm	model formula (conc ~ read_count + fraglen + GCfrac + CpGs_3)
...	other arguments to pass to bamlss

Value

   the model fit for the data

Examples

library(bamlss)
data(spike_cram_counts,package="spiky")
data(spike,package="spiky")
scc <- add_frag_info(spike_cram_counts, spike=spike)
scc$conc <- scc$conc * 0.9 # adjust for dilution
scc$CpGs_3 <- scc$CpGs ^ (1/3)
fit0 <- model_bam_standards(scc,
                            fm=conc ~ read_count + fraglen)
fit1 <- model_bam_standards(scc,
                            fm=conc ~ read_count + fraglen + GCfrac + CpGs_3)
DIC(fit0, fit1)

trichelab/spiky documentation built on Sept. 17, 2022, 8:44 a.m.