Predicting geographic location from genetic variation with deep neural networks
Most organisms are more closely related to nearby than distant members of their species, creating spatial autocorrelations in genetic data. This allows us to predict the location of origin of a genetic sample by comparing it to a set of samples of known geographic origin. Here, we describe a deep le...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
eLife Sciences Publications Ltd
2020-06-01
|
Series: | eLife |
Subjects: | |
Online Access: | https://elifesciences.org/articles/54507 |
id |
doaj-c8f96fbabc284438b9d14bd1d30b6878 |
---|---|
record_format |
Article |
spelling |
doaj-c8f96fbabc284438b9d14bd1d30b68782021-05-05T21:11:23ZengeLife Sciences Publications LtdeLife2050-084X2020-06-01910.7554/eLife.54507Predicting geographic location from genetic variation with deep neural networksCJ Battey0https://orcid.org/0000-0002-9958-4282Peter L Ralph1Andrew D Kern2https://orcid.org/0000-0003-4381-4680University of Oregon, Institute of Ecology and Evolution, Eugene, United StatesUniversity of Oregon, Institute of Ecology and Evolution, Eugene, United StatesUniversity of Oregon, Institute of Ecology and Evolution, Eugene, United StatesMost organisms are more closely related to nearby than distant members of their species, creating spatial autocorrelations in genetic data. This allows us to predict the location of origin of a genetic sample by comparing it to a set of samples of known geographic origin. Here, we describe a deep learning method, which we call Locator, to accomplish this task faster and more accurately than existing approaches. In simulations, Locator infers sample location to within 4.1 generations of dispersal and runs at least an order of magnitude faster than a recent model-based approach. We leverage Locator’s computational efficiency to predict locations separately in windows across the genome, which allows us to both quantify uncertainty and describe the mosaic ancestry and patterns of geographic mixing that characterize many populations. Applied to whole-genome sequence data from Plasmodium parasites, Anopheles mosquitoes, and global human populations, this approach yields median test errors of 16.9km, 5.7km, and 85km, respectively.https://elifesciences.org/articles/54507PlasmodiumhumanAnopheles |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
CJ Battey Peter L Ralph Andrew D Kern |
spellingShingle |
CJ Battey Peter L Ralph Andrew D Kern Predicting geographic location from genetic variation with deep neural networks eLife Plasmodium human Anopheles |
author_facet |
CJ Battey Peter L Ralph Andrew D Kern |
author_sort |
CJ Battey |
title |
Predicting geographic location from genetic variation with deep neural networks |
title_short |
Predicting geographic location from genetic variation with deep neural networks |
title_full |
Predicting geographic location from genetic variation with deep neural networks |
title_fullStr |
Predicting geographic location from genetic variation with deep neural networks |
title_full_unstemmed |
Predicting geographic location from genetic variation with deep neural networks |
title_sort |
predicting geographic location from genetic variation with deep neural networks |
publisher |
eLife Sciences Publications Ltd |
series |
eLife |
issn |
2050-084X |
publishDate |
2020-06-01 |
description |
Most organisms are more closely related to nearby than distant members of their species, creating spatial autocorrelations in genetic data. This allows us to predict the location of origin of a genetic sample by comparing it to a set of samples of known geographic origin. Here, we describe a deep learning method, which we call Locator, to accomplish this task faster and more accurately than existing approaches. In simulations, Locator infers sample location to within 4.1 generations of dispersal and runs at least an order of magnitude faster than a recent model-based approach. We leverage Locator’s computational efficiency to predict locations separately in windows across the genome, which allows us to both quantify uncertainty and describe the mosaic ancestry and patterns of geographic mixing that characterize many populations. Applied to whole-genome sequence data from Plasmodium parasites, Anopheles mosquitoes, and global human populations, this approach yields median test errors of 16.9km, 5.7km, and 85km, respectively. |
topic |
Plasmodium human Anopheles |
url |
https://elifesciences.org/articles/54507 |
work_keys_str_mv |
AT cjbattey predictinggeographiclocationfromgeneticvariationwithdeepneuralnetworks AT peterlralph predictinggeographiclocationfromgeneticvariationwithdeepneuralnetworks AT andrewdkern predictinggeographiclocationfromgeneticvariationwithdeepneuralnetworks |
_version_ |
1721458298426753024 |