Many factors have been shown to affect mating behavior. For instance, genes of the major histocompatibility complex (MHC) are known to influence mate choice in a wide variety of vertebrate species. The genetic management of captive populations can be confounded if intrinsic mate choice reduces or eliminates reproductive success between carefully chosen breeding pairs. For example, the San Diego Zoo koala colony only has a 45% copulation rate for matched individuals. Herein, we investigated determinants of koala mating success using breeding records (1984-2010) and genotypes for 52 individuals at four MHC markers. We quantified MHC diversity according to functional amino acids, heterozygosity, and the probability of producing a heterozygous offspring. We then used categorical analysis and logistic regression to investigate both copulation and parturition success. In addition, we also examined age, day length, and average pairwise kinship. Our post-hoc power analysis indicates that at a power level of 1 – β = 0.8, we should have been able to detect strong MHC preferences. However, we did not find a significant MHC effect on either copulation or parturition success with one exception: pairs with lower or no production of a joey had significantly lower MHC functional amino acid diversity in the categorical analysis. In contrast, day length and dam age (or age difference of the pair) consistently had an effect on mating success. These findings may be leveraged to improve the success of attempted pairs, conserve resources, and facilitate genetic management.
Raw_reads_fasta_files.tar.gz
Genomic DNA was extracted from whole blood or tissue samples of 52 koalas. DNA concentrations were quantified using a Nanodrop spectrophotometer and stock solutions were standardized at 5 ng/uL, with the exception of individual (ZIMS# 501277), which had a concentration of 2.5 ng/uL. We amplified primer sets previously characterized from the expressed MHC transcripts in three SDZ koalas [Abts et al. 2015], and 4 additional primer sets from Lau et al. 2013.
We began with eleven primer sets: DAA, DAB, DBA, DBB, DCB, DMA, DMB, UAB1, UB1, UE, and UK, named according to the guidelines set forth in Klein [1990]. Each set of PCR primers was optimized and used to amplify the target regions and each individual was amplified twice at each primer set in a modified Lenz-Becker approach to minimize artifact alleles. Barcoded TruSeq front-end adaptors then were added to the amplicons to uniquely identify each replicate. The products were then quantified using a Qubit fluorometer and pooled in equal concentrations of approximately 10 nM. Then the pool was sequenced using 2x300bp chemistry and an Illumina MiSeq sequencer.
The reads from each replicate were separated using the barcoded front-end adaptors. These reads are arranged by individual-replicate (indicated by the three digits - one digit after the initial six digit unique identifier). R1 or R2 represents the forward and reverse reads, respectively.
Filtered_reads_fastq_files.tar.gz
Beginning with the reads in Raw_reads_fasta_files.tar.gz, the adaptor sequences were trimmed and both ends of the read were trimmed for quality (Q = 20), using cutadapt [version 1.1b, Martin 2011]. These reads are arranged by individual-replicate (indicated by the three digits - one digit after the initial six digit unique identifier). R1 or R2 represents the forward and reverse reads, respectively.
Filtered_merged_reads_fastq_files.tar
Starting from the reads in Filtered_reads_fastq_files.tar.gz, paired reads were then merged together using pandaseq [version 2.8.1, Masella et al. 2012] and separated by amplicon using a custom perl script. The primer sets that were larger than 600 bp did not merge correctly and were not used in the final genotype sets. These reads are arranged by individual-replicate (indicated by the three digits - one digit after the initial six digit unique identifier).
Filtered_merged_amplicon_reads_fastq_files.tar.gz
AmpliSAS_Results
Barcode sequences were reincorporated into the completely filtered reads, and 1000 reads were randomly sampled from each replicate-amplicon combination for computational tractability. Preliminary genotypes were created using the AmpliSAS pipeline [Sebastian et al. 2016] with the default parameters for Illumina sequencing, except for the following: minimum dominant frequency threshold = 50%, minimum amplicon depth = 10.
Since it is thought that artifact alleles should be rare and arise from a true allele close in sequence [Babik et al. 2009], the genotypes for each amplicon were refined by filtering the alleles. For our novel primer sets, any allele that composed less than 10% of the amplicon coverage in any individual was removed. If any of these alleles always co-occurred with the same more dominant allele, the less abundant allele was removed even if it composed more than 10% of the amplicon coverage in some individuals. For the Lau et al 2013 primer sets, the AmpliSAS alleles were trimmed to match the length of those alleles presented in Lau et al 2013. Any allele that was not presented in Lau et al. 2013 was flagged and removed if it always co-occurred with the same dominant allele.
Supplementary_Material
This file contains supplementary tables and figures:
Table S1: Sequencing Summary.
Table S2: Pair Analysis Table.
Table S3: Individual Analysis Table.
Table S4: All Model Selection Table.
Table S5: Pedigree Adjustments.
Table S6: Pair Copulation Final Model.
Table S7: Male Copulation Final Model.
Table S8: Female Copulation Final Model.
Table S9: Pair Parturition Final Model.
Table S10: Male Parturition Final Model.
Table S11: Female Parturition Final Model.
Table S12: Percent variation explained by the final models.
Table S13: Permutation tests results.
Table S14: Post-hoc Power analysis of Hybrid MHC functional diversity.
Figure S1: Pair categorical analysis four different measures of MHC diversity.
Figure S2: Individual categorical analysis four different measures of internal MHC diversity.
Figure S3: Individual categorical analysis four different measures of pairwise MHC diversity.
Figure S4: MHC diversity method simulated effects of merging loci using even frequencies.
Koala_Pairings
This worksheet contains the koala pairings at the San Diego Zoo (SDZ) from 1984 to 2010. Each male and female are identified by their name, ZIMS #, and SDZ stud book number. The breeding success indicates the outcome of the pairing: 1 = offspring lives > 1 year; 2 = offspring lives < 1 year; 3 = copulation but no joey; 4 = no copulation.
Un-abridged Methods
This document contains the more detailed molecular and bioinformatic methods for those who are interested. There is some overlap with what is covered in the printed text.