Microsatellite genotyping
As published microsatellite loci for titi monkeys (Martins, 2015; Mendoza et al., 2015; Menescal, Gonçalves, Silva, Ferrari, & Schneider, 2009) revealed unreliable results for our study species, we established a new set of 27 di-repeat microsatellite loci that can be universally applied to New World monkeys (details are described in Supplementary Methods and Supplementary Tables 3–4). To simplify library preparation for genotyping by sequencing, we added adapter nucleotide sequences to the 5’ end of the locus-specific primers.
We amplified all 27 loci in single multiplex PCR reactions using the Qiagen Multiplex PCR Kit (Qiagen) with a total volume of 25 µL and containing 12.5 µL 2x Multiplex Master Mix, 1 µL of primer pool (0.2 µM of each primer), 1 µL of DNA extract (ca. 200 ng total DNA) and 10.5 µL of RNase-free water. Amplifications were performed with initial denaturation at 95°C for 15 min, 40 cycles of denaturation at 94°C for 30 sec, annealing at 57 °C for 1.5 min, extension at 72 °C for 1 min and a final extension at 60°C for 30 min. PCR products were checked on 1.5% agarose gels together with non-template controls. To prevent false homozygosity due to allelic dropout, we repeated each multiplex reaction three times (Barbian et al., 2018). In some samples, the total multiplex reaction with all 27 loci yielded low number of sequencing reads; in these cases, we additionally amplified the loci in 3 separate multiplex reactions with the following primer pools: chr01b­–chr07a, chr08a–chr12a, chr12b­–chrXa, as this method usually yielded more reads (see Supplementary Materials and Methods for details). The reactions and PCR condition for 3 separate multiplex reactions were the same as for the total multiplex reaction.
Following amplification, we pooled 5 µL of each multiplex PCR product (or of each PCR product of 3 separate multiplex reactions), purified the pooled products with the Monarch PCR & DNA Cleanup Kit (New England BioLabs) and quantified them using Qubit Fluorometer (Thermo Fisher). To prepare sequencing libraries, we performed indexing PCRs using Kapa HiFi Hotstart ReadyMix PCR Kit (Roche) with a total volume of 25 µL containing 12.5 µL 2x Kapa HiFi Hotstart ReadyMix, 1 µL (0.5 µM) of each indexing primer (containing individual barcodes) and 100 ng purified PCR product. Indexing PCRs were done with an initial denaturation step at 98°C for 45 sec, followed by 4 cycles of denaturation at 98°C for 15 sec, annealing at 62°C for 30 sec and extension at 72 °C for 30 sec, and a final extension step at 72°C for 1 min. Full-length libraries were purified with the Monarch PCR & DNA Cleanup Kit (New England BioLabs) and quantified using Qubit Fluorometer (Thermo Fisher). Fragment sizes and molarities were quantified using Bioanalyzer 2100 (Agilent Technologies). Libraries were diluted to 10 nM and then pooled and sequenced using Miseq Reagent Kit v2 with PhiX DNA (Illumina) added on the MiSeq system (Illumina). Sequencing was performed with 51 forward and 251 reverse read cycles. Only the reverse reads were used for further analysis, while forward reads were only used for MiSeq quality control.
After sequencing, the samples were demultiplexed using MiSeq Reporter software and then processed using the CHIIMP analysis pipeline (Barbian et al., 2018). The CHIIMP pipeline calls alleles by first producing unique sequences with relevant attributes (read counts, sequence length, etc.) for each MiSeq sequence file, querying the sequences for potential PCR artifacts, such as stutter sequences, and then removing all sequences that do not match the locus attributes. All alleles called by CHIIMP were manually checked to validate the results and to correct automated allele calling for those loci that contain “wobble” positions in the primer sequences and are incorrectly processed by CHIIMP.
To control for possible misidentification of animals in the field, we genotyped most individuals from 2–3 independent samples. We also used a PCR-based sexing assay (Di Fiore, 2005) to confirm reported sex (and to sex young individuals for whom sex could not be identified in the field). To control for laboratory mistakes, we genotyped each extraction twice.
Of 27 loci, 9 either consistently failed to amplify in our study animals (chr06b, chr11f, chr16b) or proved to be monomorphic (chr02a, chr02b, chr04a, chr10b, chr12a, chr13b) and were excluded from further analysis. The final set consisted of 18 loci, including 17 autosomal and one X-linked locus (chrXa) (Supplementary Table 5). All animas were genotyped at least at 10 loci (16.5 loci on average).
As X-linked loci are haploid in males, we performed all of the following summary statistics tests using data from both sexes for autosomal loci and using data only from females (diploid for X-linked loci) for locus chrXa.
We checked all loci for the presence of null alleles, allelic dropout and stuttering using Micro-Checker 2.2.5 (Van Oosterhout, Hutchinson, Wills, & Shipley, 2004). We assessed Hardy–Weinberg equilibrium (HWE) and calculated observed and expected heterozygosity with PopGenReport 2.2.2 (Adamack & Gruber, 2014). Since the presence of family structure can cause deviations from HWE and bias population genetic analyses, especially in monogamous species, we only included adults in this analysis. The analysis indicated that the population was in HWE. Two loci, chr01b and chr21a, departed from HWE, likely due to the presence of relatives in a study group and/or small sample size.
One of these two loci, chr01b, also showed evidence of null alleles. However, this locus did not show any mismatches for the known mother/offspring dyads (see below). Consequently, we ran all further analyses using two sets of data, one with the full set of loci and another one with locus chr01b excluded. Since the results from these two sets did not differ substantially, we present all further results only for the reduced data set.