areaCAT: areaCAT

In mcalgaro93/benchdamic: Benchmark of differential abundance methods on microbiome data

areaCAT

Description

Compute the area between the bisector and the concordance curve.

Usage

areaCAT(concordance, plotIt = FALSE)

Arguments

concordance	A long format data.frame produced by createConcordance function.
plotIt	Plot the concordance (default plotIt = FALSE).

Value

A long format data.frame object with several columns:

• comparison which indicates the comparison number;

• n_features which indicates the total number of taxa in the comparison dataset;

• method1 which contains the first method name;

• method2 which contains the first method name;

• rank;

• concordance which is defined as the cardinality of the intersection of the top rank elements of each list, divided by rank, i.e. , (L_{1:rank} \bigcap M_{1:rank})/(rank), where L and M represent the lists of the extracted statistics of method1 and method2 respectively;

• heightOver which is the distance between the bisector and the concordance value;

• areaOver which is the cumulative sum of the heightOver value.

See Also

createConcordance and plotConcordance

Examples

data(ps_plaque_16S)

# Balanced design for dependent samples
my_splits <- createSplits(
    object = ps_plaque_16S, varName = "HMP_BODY_SUBSITE", 
    balanced = TRUE, paired = "RSID", N = 10 # N = 100 suggested
)

# Make sure the subject ID variable is a factor
phyloseq::sample_data(ps_plaque_16S)[, "RSID"] <- as.factor(
    phyloseq::sample_data(ps_plaque_16S)[["RSID"]])

# Initialize some limma based methods
my_limma <- set_limma(design = ~ RSID + HMP_BODY_SUBSITE, 
    coef = "HMP_BODY_SUBSITESupragingival Plaque",
    norm = c("TMM", "CSS"))

# Set the normalization methods according to the DA methods
my_norm <- setNormalizations(fun = c("norm_edgeR", "norm_CSS"),
    method = c("TMM", "CSS"))

# Run methods on split datasets
results <- runSplits(split_list = my_splits, method_list = my_limma,
    normalization_list = my_norm, object = ps_plaque_16S)

# Concordance for p-values
concordance_pvalues <- createConcordance(
    object = results, slot = "pValMat", colName = "rawP", type = "pvalue"
)

# Add area over the concordance curve
concordance_area <- areaCAT(concordance = concordance_pvalues)

mcalgaro93/benchdamic documentation built on Nov. 28, 2022, 6:55 p.m.