lopo {qfasar}

Leave-one-prey-out analysis

Description

The function lopo evaluates the distinctiveness of a prey library by performing a leave-one-prey-out analysis.

Usage

lopo(sigs, type, uniq_types, type_ss, loc, dist_meas = 1, gamma = 1)

Arguments

sigs

A numeric matrix of fatty acid signatures in column-major format.

type

A character vector of prey or predator type names.

uniq_types

A character vector of the unique types, sorted alphanumerically.

type_ss

The number of signatures (sample size) for each unique type.

loc

A numeric matrix specifying the location of signatures within sigs for each uniq_types.

dist_meas

A integer indicator of the distance measure to use. Default value 1.

gamma

The power parameter of the chi-square distance measure. Default value 1.

Value

A list containing the following elements:

est

A square matrix containing the mean distribution of estimates among all prey types, by prey-type.

mean_correct

The mean proportion correctly estimated across prey types, unweighted by prey-type sample sizes.

total_correct

The proportion of all signatures correctly estimated.

n_conv

An integer vector containing the number of estimates that converged.

err_code

An integer error code (0 if no error is detected).

err_message

A string containing a brief summary of the results.

Details

The object passed as the argument sigs is intended to be the signature object returned by sig_scale or, if the prey library has been partitioned, by make_prey_part.

The objects passed as the arguments type, uniq_types, type_ss, and loc are intended to be the corresponding objects returned by prep_sig or, if the prey library has been partitioned, by make_prey_part.

The arguments dist_meas and gamma must be compatible with the function dist_between_2_sigs.

lopo performs a leave-one-prey-out analysis with a prey library and defined prey types (Bromaghin et al. 2016b). Each signature is temporarily removed from the library, the mean prey-type signature is recomputed, and the "diet" of the removed signature is estimated, after which the removed signature is returned to the library. This is done for each signature in turn. The mean estimate for each prey type is returned as a row of est. Perfect estimation would result in the square matrix est having 1.0 along its diagonal and 0.0 in all off-diagonal positions. Large off-diagonal elements are indicative of confounding, or similarity, between the corresponding prey types. The returned object mean_correct is the mean of the diagonal elements of est, while the returned object total_correct is the mean computed over all signatures in the prey library.

Note: the statistics are computed based on the estimates that successfully converge (n_conv) and prey types that only have a sample size of 1 are skipped.

Because of the numerical optimization involved in a leave-one-prey-out analysis, lopo can take a few minutes to run with a large prey library.

The statistics computed by lopo are one measure of the distinctiveness of prey types within a prey library. However, it is important to be aware that such statistics are not necessarily informative of the ability of QFASA to accurately estimate predator diets, as Bromaghin et al. (2015, 2016a, 2016b) found that QFASA performance depends strongly on the interaction between characteristics of a prey library, the specific diet of a predator, and the accuracy of the calibration coefficients. Consequently, the user is warned not to misinterpret or misrepresent these statistics.

References

Bromaghin, J.F., S.M. Budge, and G.W. Thiemann. 2016b. Should fatty acid signature proportions sum to 1 for diet estimation? Ecological Research 31:597-606.

Bromaghin, J.F., S.M. Budge, G.W. Thiemann, and K.D. Rode. 2016b. Assessing the robustness of quantitative fatty acid signature analysis to assumption violations. Methods in Ecology and Evolution 7:51-59.

Bromaghin, J.F., K.D. Rode, S.M. Budge, and G.W. Thiemann. 2015. Distance measures and optimization spaces in quantitative fatty acid signature analysis. Ecology and Evolution 5:1249-1262.

Examples

lopo(sigs = matrix(c(0.05, 0.10, 0.30, 0.55,
                     0.04, 0.11, 0.29, 0.56,
                     0.10, 0.05, 0.35, 0.50,
                     0.12, 0.03, 0.37, 0.48,
                     0.10, 0.06, 0.35, 0.49,
                     0.05, 0.15, 0.35, 0.45), ncol=6),
     type = c("Type_1", "Type_1", "Type_2", "Type_2", "Type_3", "Type_3"),
     uniq_types = c("Type_1", "Type_2", "Type_3"),
     type_ss <- c(2, 2, 2),
     loc = matrix(c(1, 3, 5, 2, 4, 6), ncol=2),
     dist_meas = 1)

lopo(sigs = matrix(c(0.05, 0.10, 0.30, 0.55,
                     0.04, 0.11, 0.29, 0.56,
                     0.10, 0.05, 0.35, 0.50,
                     0.12, 0.03, 0.37, 0.48,
                     0.10, 0.06, 0.35, 0.49,
                     0.05, 0.15, 0.35, 0.45), ncol=6),
     type = c("Type_1", "Type_1", "Type_2", "Type_2", "Type_3", "Type_3"),
     uniq_types = c("Type_1", "Type_2", "Type_3"),
     type_ss <- c(2, 2, 2),
     loc = matrix(c(1, 3, 5, 2, 4, 6), ncol=2),
     dist_meas = 2)

lopo(sigs = matrix(c(0.05, 0.10, 0.30, 0.55,
                     0.04, 0.11, 0.29, 0.56,
                     0.10, 0.05, 0.35, 0.50,
                     0.12, 0.03, 0.37, 0.48,
                     0.10, 0.06, 0.35, 0.49,
                     0.05, 0.15, 0.35, 0.45), ncol=6),
     type = c("Type_1", "Type_1", "Type_2", "Type_2", "Type_3", "Type_3"),
     uniq_types = c("Type_1", "Type_2", "Type_3"),
     type_ss <- c(2, 2, 2),
     loc = matrix(c(1, 3, 5, 2, 4, 6), ncol=2),
     dist_meas = 3,
     gamma = 0.25)

lopo(sigs = matrix(c(0.05, 0.10, 0.30, 0.55,
                     0.04, 0.11, 0.29, 0.56,
                     0.10, 0.05, 0.35, 0.50,
                     0.12, 0.03, 0.37, 0.48,
                     0.10, 0.06, 0.35, 0.49,
                     0.05, 0.15, 0.35, 0.45), ncol=6),
     type = c("Type_1", "Type_1", "Type_2", "Type_2", "Type_3", "Type_3"),
     uniq_types = c("Type_1", "Type_2", "Type_3"),
     type_ss <- c(2, 2, 2),
     loc = matrix(c(1, 3, 5, 2, 4, 6), ncol=2),
     dist_meas = 3)

lopo(sigs = matrix(c(0.05, 0.10, 0.30, 0.55,
                     0.04, 0.11, 0.29, 0.56,
                     0.10, 0.05, 0.35, 0.50,
                     0.12, 0.03, 0.37, 0.48,
                     0.10, 0.06, 0.35, 0.49,
                     0.05, 0.15, 0.35, 0.45), ncol=6),
     type = c("Type_1", "Type_1", "Type_2", "Type_2", "Type_3", "Type_3"),
     uniq_types = c("Type_1", "Type_2", "Type_3"),
     type_ss <- c(2, 2, 2),
     loc = matrix(c(1, 3, 5, 2, 4, 6), ncol=2))