Table 3.

Model selection results from demographic inference using the site-frequency spectrum (SFS) in fastsimcoal2

PopulationModelNeCONNeANCNeBOTTBOTΔLAICΔAIC
AlpsBOT128,11782,32181219337,862249,4950
HarzBOT35589,84864,104153578,191513,2860
NortheastBOT74965,46448,177262681,403484,9180
NorthwestBOT124157,89153,7814910141,4721,160,7150
AlpsISC915847,203NANA37,882249,59195
HarzISC135179,897NANA78,194513,30519
NortheastISC231824,407NANA81,403484,9212
NorthwestISC5931112,707NANA141,4811,160,76247
AlpsNull26,348NANANA37,961249,947452
HarzNull34,143NANANA78,713515,6882402
NortheastNull50,657NANANA81,587485,760841
NorthwestNull90,174NANANA142,1921,164,0263311

[i] Three models were run for each regional grouping of Alps, Harz, Northeast, and Northwest (population). Model refers to one of three model types: Null model of constant population size (Null), instantaneous size change/bottleneck (BOT), and ISC. Parameters were estimated for contemporary Ne (NeCON), ancestral Ne (NeANC), Ne during bottleneck (NeBOT), and time of bottleneck in number of generations (TBOT). Model fit and selection criterion are shown as delta likelihood, calculated as maximum estimated likelihood maximum observed likelihood (ΔL), AIC, and ΔAIC.