Many population genetics tools employ composite likelihoods, because fully modeling genomic linkage is challenging. But traditional approaches to estimating parameter uncertainties and performing model selection require full likelihoods, so these tools have relied on computationally expensive maximum-likelihood estimation (MLE) on bootstrapped data. Here, we demonstrate that statistical theory can be applied to adjust composite likelihoods and perform robust computationally efficient statistical inference in two demographic inference tools: ∂a∂i and TRACTS. On both simulated and real data, the adjustments perform comparably to MLE bootstrapping while using orders of magnitude less computational time.
%0 Journal Article
%1 coffman2016computationally
%A Coffman, Alec J.
%A Hsieh, Ping Hsun
%A Gravel, Simon
%A Gutenkunst, Ryan N.
%D 2016
%J Molecular Biology and Evolution
%K AIC bootstrap composite_likelihood confidence_interval demographic_inference information_matrix likelihood methods statistics
%N 2
%P 591-593
%R 10.1093/molbev/msv255
%T Computationally Efficient Composite Likelihood Statistics for Demographic Inference
%U http://mbe.oxfordjournals.org/content/33/2/591.abstract
%V 33
%X Many population genetics tools employ composite likelihoods, because fully modeling genomic linkage is challenging. But traditional approaches to estimating parameter uncertainties and performing model selection require full likelihoods, so these tools have relied on computationally expensive maximum-likelihood estimation (MLE) on bootstrapped data. Here, we demonstrate that statistical theory can be applied to adjust composite likelihoods and perform robust computationally efficient statistical inference in two demographic inference tools: ∂a∂i and TRACTS. On both simulated and real data, the adjustments perform comparably to MLE bootstrapping while using orders of magnitude less computational time.
@article{coffman2016computationally,
abstract = {Many population genetics tools employ composite likelihoods, because fully modeling genomic linkage is challenging. But traditional approaches to estimating parameter uncertainties and performing model selection require full likelihoods, so these tools have relied on computationally expensive maximum-likelihood estimation (MLE) on bootstrapped data. Here, we demonstrate that statistical theory can be applied to adjust composite likelihoods and perform robust computationally efficient statistical inference in two demographic inference tools: ∂a∂i and TRACTS. On both simulated and real data, the adjustments perform comparably to MLE bootstrapping while using orders of magnitude less computational time.},
added-at = {2016-05-02T22:07:00.000+0200},
author = {Coffman, Alec J. and Hsieh, Ping Hsun and Gravel, Simon and Gutenkunst, Ryan N.},
biburl = {https://www.bibsonomy.org/bibtex/2704e8fb1cc0fc23d8e198b3eb274c9f4/peter.ralph},
doi = {10.1093/molbev/msv255},
eprint = {http://mbe.oxfordjournals.org/content/33/2/591.full.pdf+html},
interhash = {1170e329b599dbb52af7bb6c4a317d8a},
intrahash = {704e8fb1cc0fc23d8e198b3eb274c9f4},
journal = {Molecular Biology and Evolution},
keywords = {AIC bootstrap composite_likelihood confidence_interval demographic_inference information_matrix likelihood methods statistics},
number = 2,
pages = {591-593},
timestamp = {2016-05-02T22:16:42.000+0200},
title = {Computationally Efficient Composite Likelihood Statistics for Demographic Inference},
url = {http://mbe.oxfordjournals.org/content/33/2/591.abstract},
volume = 33,
year = 2016
}