Detect contaminants by correlating taxon reads with total sample depth

Flags potential contaminant taxa from the correlation between each taxon's relative abundance and total sample depth. Contaminants, introduced at a roughly constant level during processing, show a negative correlation (their proportion is inflated in low-biomass samples). Inspired by the GRIMER tool.

This is a pure detector: it flags taxa and returns a contam_tbl without modifying physeq. Use filter_contam_pq() to remove the flagged taxa and plot.contam_tbl() for a quick diagnostic.

Usage

identify_contam_corr_pq(
  physeq,
  min_total_reads = 1000,
  min_prevalence = 2,
  contam_threshold = -0.5,
  verbose = TRUE
)

Arguments

physeq

(phyloseq, required) A phyloseq object.

min_total_reads

(numeric, default 1000) Samples with fewer total reads are excluded to avoid spurious correlations from low-depth outliers.

min_prevalence

(integer, default 2) Taxa present in fewer than this many samples are excluded from the analysis.

contam_threshold

(numeric, default -0.5) Taxa whose relative- abundance correlation (cor_rel) is below this value are flagged. Use values closer to 0 to flag more aggressively.

verbose

(logical, default TRUE) Print a one-line detection summary.

Value

A contam_tbl with one row per flagged taxon, method = "corr", and the evidence columns total_reads, prevalence, cor_abs, cor_rel, slope, and p_value. Empty when nothing is flagged.

See also

filter_contam_pq() to remove flagged taxa, contam_tbl for the return type.

Author

Adrien Taudière

Examples

library(MiscMetabar)
# \donttest{
# Inject a constant-reads fake contaminant so the example flags something
otu <- as(phyloseq::otu_table(data_fungi), "matrix")
fake <- matrix(rep(50, nrow(otu)), ncol = 1,
  dimnames = list(rownames(otu), "fake_contaminant"))
tax <- rbind(
  as.matrix(phyloseq::tax_table(data_fungi)),
  fake_contaminant = rep("Contaminant", ncol(phyloseq::tax_table(data_fungi)))
)
data_fungi_c <- phyloseq::phyloseq(
  phyloseq::otu_table(cbind(otu, fake), taxa_are_rows = FALSE),
  phyloseq::tax_table(tax),
  phyloseq::sample_data(data_fungi)
)

flagged <- identify_contam_corr_pq(data_fungi_c)
#> ℹ corr: flagged 1 of 1200 taxa (0.1%) at cor_rel < -0.5.
flagged
#> contam_tbl: 1 flagged taxon from method "corr"
#> # A tibble: 1 × 8
#>   taxon          method total_reads prevalence cor_abs cor_rel    slope  p_value
#>   <chr>          <chr>        <dbl>      <dbl>   <dbl>   <dbl>    <dbl>    <dbl>
#> 1 fake_contamin… corr          7150        143      NA  -0.667 -6.32e-7 8.96e-20
# }