DSFT: Downsampling followed by fringe trimming

In QSutils: Quasispecies Diversity

Description

Diversity indices are influenced to a greater or lesser degree by the sample size on which they are computed. This function helps to minimize the bias inherent to sample size. First the vector of abundances is scaled to a smaller sample size, then all haplotypes with abundances below a given threshold are excluded with high confidence.

Usage

DSFT(nr, size, p.cut = 0.002, conf = 0.95)

Arguments

nr	Vector of observed haplotype counts.
size	Size to downsample.
p.cut	Abundance threshold.
conf	Confidence in trimming.

Value

Vector of logicals, with false the haplotypes to be excluded.

Author(s)

Mercedes Guerrero-Murillo and Josep Gregori

References

Gregori J, Perales C, Rodriguez-Frias F, Esteban JI, Quer J, Domingo E. Viral quasispecies complexity measures. Virology. 2016 Jun;493:227-37. doi: 10.1016/j.virol.2016.03.017. Epub 2016 Apr 6. Review. PubMed PMID: 27060566.

Gregori J, Salicrú M, Domingo E, Sanchez A, Esteban JI, Rodríguez-Frías F, Quer J. Inference with viral quasispecies diversity indices: clonal and NGS approaches. Bioinformatics. 2014 Apr 15;30(8):1104-1111. Epub 2014 Jan 2. PubMed PMID: 24389655.

Examples

# Generate viral quasispecies abundance data.
set.seed(123)
n <- 2000
y <- geom.series(n,0.8)+geom.series(n,0.0004)
nr.pop <- round(y*1e7)
# Get a sample of 10000 reads from this population.
sz2 <- 10000
nr.sz2 <- table(sample(length(nr.pop),size=sz2,replace=TRUE,p=nr.pop))
# Filter out haplotypes below 0.1%.
thr <- 0.1
fl <-  nr.sz2>=sz2*thr/100
nr.sz2 <- nr.sz2[fl]
Shannon(nr.sz2) #0.630521
# DSFT to 5000 reads.
sz1 <- 5000
fl <- DSFT(nr.sz2,sz1)
nr.sz2 <- nr.sz2[fl]
# Compute size corrected Shannon entropy.
Shannon(nr.sz2) #0.6189798

QSutils documentation built on Nov. 8, 2020, 7:42 p.m.