Network-based hierarchical population structure analysis for large genomic data sets

Abstract

Analysis of population structure in natural populations using genetic data is a common practice in ecological and evolutionary studies. With large genomic data sets of populations now appearing more frequently across the taxonomic spectrum, it is becoming increasingly possible to reveal many hierarchical levels of structure, including fine-scale genetic clusters. To analyze these data sets, methods need to be appropriately suited to the challenges of extracting multilevel structure from whole-genome data. Here, we present a network-based approach for constructing population structure representations from genetic data. The use of community-detection algorithms from network theory generates a natural hierarchical perspective on the representation that the method produces. The method is computationally efficient, and it requires relatively few assumptions regarding the biological processes that underlie the data. We show the approach by analyzing population structure in the model plant species Arabidopsis thaliana and in human populations. These examples illustrate how networkbased approaches for population structure analysis are well-suited to extracting valuable ecological and evolutionary information in the era of large genomic data sets.

Population genomics of postglacial western eurasia

“…To obtain a finer-scale characterization of genetic ancestries across space and time, we assigned imputed ancient individuals to genetic clusters by applying hierarchical community detection on a network of pairwise identity-by-descent (IBD)-sharing similarities 28 (Extended Data Fig. 3; Supplementary Note 3c).…”

“…These results are consistent with much higher genetic differentiation between ancient Europeans than present-day populations reflecting lower effective population sizes and genetic isolation among ancient groups. To obtain a finer-scale characterization of genetic ancestries across space and time, we assigned imputed ancient individuals to genetic clusters by applying hierarchical community detection on a network of pairwise identity-by-descent (IBD)-sharing similarities 28 (Extended Data Fig. 3; Supplementary Note 3c).…”

Extracting hierarchical features of cultural variation using network-based clustering

“…In the second step of NetStruct , we iteratively remove edges with lower weights from the network to reveal the finer-scale structure within coarser communities. NetStruct uses a community-detection Louvain algorithm ( Blondel et al, 2008 ) together with an iterative edge-pruning method ( Greenbaum et al, 2019 ). The Louvain algorithm maximizes a ‘modularity score’ for each community, quantifying the difference between the actual density of edges within the community and the expected density if all edges in the network were distributed at random while preserving the degree distribution of the network.…”

“…For a given dataset of interest, the similarity function is chosen in a manner suited to the application. We follow Greenbaum et al (2016Greenbaum et al ( , 2019 in choosing frequency-weighted similarity measures that emphasize shared rare values of a character.…”

“…Owing to the importance of shared descent in influencing cultural variation, tools for analysing cultural variation data can employ methods suited to the analysis of genetic data, which also possess signatures of shared descent; thus, many statistical methods used in cultural data analysis are similar to those used for genetic data ( Bromham, 2017 ; Gray et al, 2010 ; Pagel, 2009 ). Recently, we have introduced a method, NetStruct , for use in understanding genetic variation data that result from hierarchical genetic structure ( Greenbaum et al, 2016 , 2019 ). The method, employing ideas from network analysis, produces a distinctive form of visualization of hierarchical population relationships.…”