Samples
Dr. Heidi Parker (NHGRI, NIH, Bethesda, MD) provided de-identified and breed-verified samples from forty dogs including ten Shi Tzu, ten Dachshund, ten Newfoundland, and ten Irish Wolfhounds. All ethical approvals for protocols, including IACUC approval for collection of the samples, were obtained for this study.
TERT sequence
We used the genomic sequence of canine TERT (TERT_CANFA ENSCAFT00000017081 CanFam 2.0) [Release 58 May 2010 [12]] as a reference, including 500 additional nucleotides flanking either end of the gene. The plasmid editing program, ApE (Wayne Davis; Salt Lake City, Utah; version “ApE_OSX_1_1_7.dmg”) [13] was used to parcel exons, introns, and previously catalogued polymorphisms. As depicted in Additional file 1: Figure S1, the gene was divided into 500 nucleotide-long segments, with each segment overlapping by 250 nucleotides, for a total of 74 sections. The second primer pair could not be generated due to extensive GC pairing. ApE was used to generate forward and reverse primers for each section. We then manually reviewed each list of potential primers per section, with the final primer pairs chosen based on stability, GC content, melting point, and salt content. The primer sequences are listed in Additional file 1: Table S1.
DNA extraction
We extracted DNA from lymph node tissue of a healthy dog obtained with IACUC approval, and used it at a concentration of 40 ng/μL to standardize the conditions of PCR amplification. The performance of each pair was verified by conventional PCR and gel electrophoresis. Of 69 primer pairs that successfully amplified DNA, we chose 11 that were interspersed throughout the gene sequence. The amplified DNA was then sequenced in both forward and reverse directions for a total of 22 samples. These partial DNA sequences were examined for length, quality, and how closely they matched the reference canine TERT sequence.
PCR amplification
Two sets of twenty dogs were used to derive the sequence: the first set was used to run primer pairs 1 through 40 and 66 through 75, and the second set was used to run primer pairs 41 through 65. A DNA concentration of 20 ng/μL was used for each sample for PCR, with go-Taq polymerase (Promega). PCR was performed at 42-55°C for 40 cycles. Each PCR was verified by carrying out gel electrophoresis of 8 μL of sample in a 2% agarose gel. The remaining 12 μL of successful samples were subjected to capillary sequencing (Biomedical Genomic Center, University of Minnesota).
Sequence analysis and alignment
We compiled the resulting sequences using CodonCode Aligner (CodonCode; Dedham, Massachusetts) [14] with default parameters (Version 3.7.1.1), aligning each amplicon to the published reference sequence. The consensus sequences of each dog were aligned using clustalx (Version 2.0.11) [15]. We considered three or more identical variants from the original reference sequence to be polymorphisms. At this point each group of 20 dogs was handled separately. Gaps and ambiguous alignment segments were trimmed in Jalview4 (Version 11.0) [16] prior to input into three complementary programs for phylogenetic tree determination: 1. MrBayes (Version 3.2) [17], a Bayesian inference algorithm that utilizes Markov Chain Monte Carlo optimization (parameters: default General Time Reversible model of evolution, invariant sites, gamma distributed rates, 2 million generations) 2. Simple DNA parsimony as implemented in dnapars using consense (Majority Rule consensus) from the Phylip package (Version 3.69) [18]. Neighbor Joining algorithm implemented in clustalx. We visualized trees output by the tree-building programs using FigTree [19].