Abstract Detail


Hoffberg, Sandra [1], Eaton, Deren [2].

Amount of hybridization differs with varying gene tree lengths in reference aligned RADseq data.

Phylogenetic methods that model the multispecies coalescent (in ASTRAL) or network multispecies coalescent (in PhyloNetworks and MSCquartets) require gene trees as input.  Ideally, all bases within a locus used for inferring a gene tree would have the same evolutionary history. However, the breakpoints that separate different genealogical histories are not able to be directly observed, and several might fall within a sampled region of the chromosome.  Because this information is unknown, it is unclear what size genomic window should be used.  We apply a new ipyrad tool, treeslider, which cuts output files from sample aligned reference-based assemblies into separate loci. These windows can be any size (up to the scaffold length) and can overlap by any amount. The extracted loci can then be used as loci for species tree inference. Using RADseq data from over 200 samples from 37 species in the genus Amaranthus, we assessed hybridization between species using gene trees inferred from different locus lengths and inference in ASTRAL and nanuq. We found that as the number of base pairs used for gene tree inferences decreases from 7% of the average chromosome length to less than 1% of the average chromosome length, additional introgression can be detected at first, until the loci become so small that gene tree estimation error becomes too great. We also test the amount of hybridization detected sequence data generated using the coalescent simulation package ipcoal. As systematic research facilitated by a reference genome becomes more common, methods like treeslider will enable researchers to study spatial gene tree variation using reduced representation genetic data.   

1 - Columbia University, E3B, 1200 Amsterdam Ave, Schermerhorn Ext, 10th Floor, New York, NY, 10027, United States
2 - Columbia University, Ecology, Evolution, And Environmental Biology, 1200 Amsterdam Ave. , Schermerhorn Ext. Office 1007, New York, NY, 10027, United States

Gene tree discordance
simulated sequence data
network analysis

Presentation Type: Oral Paper
Session: PHYL3, Phylogenomics III
Location: Virtual/Virtual
Date: Friday, July 31st, 2020
Time: 10:15 AM
Number: PHYL3002
Abstract ID:644
Candidate for Awards:None

Copyright © 2000-2020, Botanical Society of America. All rights reserved