runGenphen: Genetic association analysis using Bayesian inference and...

In snaketron/genphen: genphen: tool for quantification of genotype-phenotype associations in genome wide association studies (GWAS)

Description

Given a set of genotypes (single nucleotide polymorphisms - SNPs; or single amino acid polymorphisms - SAAPs) for a set of individuals, and a corresponding set of phenotypes, genphen quantifies the association between each genotype and phenotype using Bayesian inference and statistical learning.

Usage

runGenphen(genotype, phenotype, phenotype.type, model.type,
           mcmc.chains, mcmc.steps, mcmc.warmup, cores, 
           hdi.level, stat.learn.method, cv.steps, ...)

Arguments

genotype	Character matrix/data frame or a vector, containing SNPs/SAAPs as columns or alternatively as DNAMultipleAlignment or AAMultipleAlignment Biostrings object.
phenotype	Numerical vector (for a single phenotype) or matrix with multiple phenotypes stored as columns.
phenotype.type	Vector representing the type of each phenotype (of the phenotype input), with 'Q' identifier for quantitative, or 'D' for dichotomous phenotypes.
model.type	Type of Bayesian model: 'univariate' or 'hierarchical'
mcmc.chains	Number of MCMC chains (default = 2).
mcmc.steps	Length of MCMC chains (default = 1,000).
mcmc.warmup	Length of adaptive part of MCMC chains (default = 500).
cores	Number of cores to use (default = 1).
hdi.level	Highest density interval (HDI) (default = 0.95).
stat.learn.method	Parameter used to specify the statistical learning method used in the analysis. Currently two methods are available: random forest ('rf') and support vector machine ('svm'). For no statistical learning select 'none'.
cv.steps	cross-validation steps (default = 1,000).
...	Optional parameters include adapt_delta: STAN configuration (default = 0.9); max_treedepth: STAN configuration (default = 10); ntree: Number of random forest trees to grow, only in case stat.learn.method = 'rf' (default = 1000); cv.fold: Cross-validation fold (default = 0.66).

Details

Input:

• genotype genotype data (e.g. set of 1,000 SNPs found along the aligned genomes of 10 individuals) - provided in one of three possible input types:

  • character vector of length N (if only a single SNP/SAAP is provided), containing the genotypes of N individuals.

  • character matrix with dimensions NxS (N = individuals, S = SNPs/SAAPs).

  • AAMultipleAlignment or DNAMultipleAlignment object; if the genotype data is a multiple sequence alignment composed of N sequences.

• phenotype phenotype data (dichotomous or quantitative phenotypes allowed)

  • numerical vector of length N if only a single phenotype is analyzed

  • numerical matrix NxP, if P phenotypes are provided.

• phenotype.type Vector with identifiers specifying the type of the phenotypes with 'Q' (for quantitative) or 'D' (for dichotomous) for each column in the phenotype dataset.

• model.type Specifies the structure of Bayesian model used to estimate the effect size of each genotype. Options allow for either 'univariate' (each SNP/SAAP treated as completely independent) or 'hierarchical' (SNP/SAAP effects share information through partial pooling).

Metrics: To quantify the association between each genotype and phenotype genphen computes multiple measures of association:

• Effect size (beta): for each SNP we compute beta (effect) with Bayesian inference). beta quantifies the strength of the association between the genotypes and the phenotype. We report for each beta its mean and 95% (for instance) highest density interval (HDI) of beta, which is defined as the interval that covers a 95% of the posterior distribution, with every point inside the interval having a higher credibility than any point outside it.

• Classification accuracy (CA): CA measures the degree of accuracy with which one can classify (predict) the alleles of a SNP from the phenotype. If there exists a strong association between a particular SNP and the phenotype, one should be able to train a statistical model (using RF or SVM) which accurately classifies the two alleles of that SNP solely from the phenotype data (CA close to 1). Otherwise, the model should perform poorly, with the classification accuracy of the model being approximately similar to that of simple guessing (CA close to 0.5)

• Cohen's kappa statistic: There is one pitfall where the CA estimate can be misleading, and this is the case when the analyzed SNP is composed of unevenly represented genetic states (alleles). For instance, the allele A of a given SNP is found in 90% of the individuals, while the other allele T in only 10%. Such an uneven composition of the alleles can lead to misleading results, i.e. even without proper learning the algorithm can produce a high $CA close to 0.9 simply by always predicting the dominant label. The kappa statistics is a quality metric, which is to be used together with CA. Cohen defines the following meaningful kappa intervals: [kappa<0]: “no agreement”, [0.0-0.2]: “slight agreement”, [0.2-0.4]: “fair agreement”, [0.4-0.6]: “moderate agreement”, [0.6-0.8]: “substantial agreement” and [0.8-1.0]: “almost perfect agreement”.

Value

General parameters:

site	id of the site (e.g. position in the provided sequence alignment)
ref, alt	reference and alternative genotypey
refN, altN	count of ref and alt genotypes
phenotype.id	Identifier of the studied phenotype

Association scores:

beta.mean, beta.se, beta.sd, beta.hdi.low/beta.hdi.high	Estimates of the mean, standard error, standard deviation and HDI of the slope coefficient
ca.mean, ca.hdi.low/ca.hdi.high	CA estimate and HDI
kappa.mean, kappa.hdi.low/kappa.hdi.high	Cohen's kappa and HDI
rank	Pareto optimiazion based front (rank) of SNP/SAAP estimated by maximizing metrics beta.mean and kappa.mean

MCMC convergence parameters:

Neff	Effective sampling size
Rhat	Potential scale reduction factor

Posterior predictions:

ppc	Posterior prediction check and real data summary for each genotype.

Posterior summary:

complete.posterior

Complete stan object containing the posterior of each parameter estimated during the Bayesian inference. The data can be used for model debugging, posterior predictive checks, etc.

Author(s)

Simo Kitanovski <[e-mail]>

See Also

runDiagnostics, runPhyloBiasCheck

Examples

# genotypes:
data(genotype.saap)
# quantitative phenotype:
data(phenotype.saap)
# dichotomous phenotype:
data(dichotomous.phenotype.saap)
# make phenotype matrix (column = phenotype)
phenotypes <- cbind(phenotype.saap, dichotomous.phenotype.saap)

# run genphen
out <- runGenphen(genotype = genotype.saap[, 80:82],
                  phenotype = phenotypes,
                  phenotype.type = c("Q", "D"),
                  model.type = "univariate",
                  mcmc.chains = 4,
                  mcmc.steps = 1500,
                  mcmc.warmup = 500,
                  cores = 1,
                  hdi.level = 0.95,
                  stat.learn.method = "rf",
                  cv.steps = 200)

snaketron/genphen documentation built on Oct. 1, 2021, 8:09 a.m.