Genetic analysis
DNA extraction and individual identification followed the protocol of Poutanen et al. (2019). The only modifications for the microsatellite PCR protocol were that final concentration of primer Rt5 was decreased from 0.2 μmol/l to 0.1 μmol/l and BSA concentration from 0.1 μg/μl to 0.01 μg/μl. We genotyped samples using 14 microsatellite markers and used the multitube approach by performing three PCR replicates for each sample to minimize genotyping errors (Taberlet et al. 1996). Samples which were successfully amplified with at least 11 loci were used to establish individual identity. Consensus genotypes were constructed with a rule that alleles of homozygous loci were needed to amplify three times and of heterozygous loci two times in order to be accepted as final genotype (Jansson et al. 2014, Stansbury et al. 2014). When matching genotypes, we allowed two mismatches in different loci between the samples in order to accept them as representing the same individual. Thus, we required at least six matching loci between the samples to assign them to the same individual. This was a sufficient number of loci for identification as when counting the probability of identity (PI) values based on the six most uninformative loci the PI values were in the range of recommendations (PI=0.0003 in 2016 and PI=0.0001 in 2017, PIDsib=0.02 in both years) (Waits et al. 2001). Consensus genotypes were matched to individuals using the software Cervus v. 3.0.7 (Kalinowski et al. 2007) and Gimlet v. 1.3.3 (Valière 2002). At least one DNA sample of each identified individual were sexed with X- and Y- chromosome specific primer pair ZFX/ZFY following the protocol of Poutanen et al. 2019.
We used the consensus genotype data to calculate deviations from Hardy Weinberg Equilibrium using Genepop v. 4.2 (Rousset 2008) and number of alleles and expected and observed heterozygosities using Gimlet. Poutanen et al. (2019) demonstrated that fecal sampling risks mistaking target and non-target species (roe deer, Capreolus capreolus ), but demonstrated that the only roe deer samples were those where seven or less microsatellites amplified. We therefore assumed that the 12 to 14 microsatellite markers used here were sufficient to exclude roe deer DNA and did not perform any additional species identification testing.