summix: summix

In hendriau/Summix: Summix2: A suite of methods to estimate, adjust, and leverage substructure in genetic summary data

summix

Description

Estimating mixture proportions of reference groups from large (N SNPs>10,000) genetic AF data.

Usage

summix(
  data,
  reference,
  observed,
  pi.start = NA,
  goodness.of.fit = TRUE,
  override_removeSmallRef = FALSE,
  network = FALSE,
  N_reference = NA,
  reference_colors = NA
)

Arguments

data	A dataframe of the observed and reference allele frequencies for N genetic variants. See data formatting document at https://github.com/hendriau/Summix for more information.
reference	A character vector of the column names for the reference groups.
observed	A character value that is the column name for the observed group.
pi.start	Length K numeric vector of the starting guess for the reference group proportions. If not specified, this defaults to 1/K where K is the number of reference groups.
goodness.of.fit	Default value is TRUE. If set as FALSE, the user will override the default goodness of fit measure and return the raw objective loss from slsqp.
override_removeSmallRef	Default value is FALSE. If set as TRUE, the user will override the automatic removal of reference groups with <1% global proportions - this is not recommended.
network	Default value is FALSE. If set as TRUE, function will return a network diagram with nodes as estimated substructure proportions and edges as degree of similarity between the given node pair.
N_reference	numeric vector of the sample sizes for each of the K reference groups; must be specified if network = "TRUE".
reference_colors	A character vector of length K that specifies the color each reference group node in the network plot. If not specified, this defaults to K random colors.

Value

A data frame with the following columns:

goodness.of.fit: scaled objective loss from slsqp() reflecting the fit of the reference data. Values between 0.5-1.5 are considered moderate fit and should be used with caution. Values greater than 1.5 indicate poor fit, and users should not perform further analyses using Summix.

iterations: number of iterations for SLSQP algorithm

time: time in seconds of SLSQP algorithm

filtered: number of genetic variants not used in the reference group mixture proportion estimation due to missing values.

K columns of mixture proportions of reference groups input into the function

Author(s)

Adelle Price, [e-mail]

Hayley Wolff, [e-mail]

Audrey Hendricks, [e-mail]

References

https://github.com/hendriau/Summix

See Also

https://github.com/hendriau/Summix for further documentation and https://github.com/hendriau/Summix2_manuscript for a larger sample data set and description of simulations in Summix2 manuscript. slsqp function in the nloptr package for further details on Sequential Quadratic Programming https://www.rdocumentation.org/packages/nloptr/versions/1.2.2.2/topics/slsqp

Examples

# load the data
data("ancestryData")

# Estimate 5 reference ancestry proportion values for the gnomAD African/African American group
# using a starting guess of .2 for each ancestry proportion.
summix(data = ancestryData,
    reference=c("reference_AF_afr",
        "reference_AF_eas",
        "reference_AF_eur",
        "reference_AF_iam",
        "reference_AF_sas"),
    observed="gnomad_AF_afr",
    pi.start = c(.2, .2, .2, .2, .2),
    goodness.of.fit=TRUE)

hendriau/Summix documentation built on Nov. 13, 2024, 6:53 a.m.