Telomere length measurements
Relative erythrocyte telomere lengths (TL) of 2746 nestlings from Hestmannøy and Træna were successfully measured using the real-time quantitative polymerase chain reaction (qPCR) amplification method by Cawthon (2002) with modifications by Criscuolo et al. (2009). Primer sequences, PCR assay setup and thermal profiles followed Pepke et al. (2021, submitted ) and are detailed in Appendix S1. Briefly, this method measures the ratio of telomere sequence relative to the amount of a non-variable gene (GAPDH) and a reference sample. The reference sample consisted of pooled DNA from 6 individuals, which was also included as a 2-fold serial dilution (40-2.5 ng/well) on all plates to produce a standard curve, in addition to a non-target control sample (all in triplicates). Samples were randomized and run on 2x127 96-well plates (telomere and GAPDH assays, respectively). The qPCR data was analyzed using the qBASE software (Hellemans, Mortier, De Paepe, Speleman, & Vandesompele, 2007), which computes relative TL as the ratio (T/S) of the telomere repeat copy number (T) to a single copy gene number (S) similar to Cawthon (2002). In qBASE the T/S ratio is calculated as calibrated normalized relative quantities (CNRQ) that control for differences in amplification efficiency between plates and for inter-run variation by including three inter-run calibrators from the standard curve. All individual plate efficiencies were within 100±10% (mean telomere assay efficiency was 97.5±3.9%, and 97.6±4.2% for GAPDH assays). The average of the reference sample cycle thresholds (Ct) across all plates were 10.54±0.03 S.D. and 21.53±0.02 S.D. for telomere and GAPDH assays, respectively. Thus, while reproducibility of TL measurements within the reference sample of the same DNA sample extract is high, we performed DNA re-extraction of the same blood samples for 25 individuals to test TL consistency across DNA extractions (Appendix S1). The re-extractions were run on different plates and the TL estimates of these samples remained highly correlated (R2=0.75, Fig. S1.2). For these individuals, the average of the TL measurements was used in subsequent analyses. All reactions for the primary analyses (from the populations on Hestmannøy and Træna) were performed by the same person (MLP). MLP and WB generated the secondary dataset (n =569 on 2x21 plates, from the populations on Leka and Vega) as described in Pepke et al. (2021, submitted ). The primary and secondary datasets used different reference samples and are therefore not combined in the analyses.