DNA methylation measurements
The MS-AFLP protocol used modifies the standard AFLP protocol by substituting the MseI enzyme with the methylation-sensitive isoschizomeric enzymes MspI and HpaII (New England Biolabs). See Salmon et al., (2008) for further details on the MS-AFLP method. Together, four types of variation can be scored: Type 1 is when both enzymes cut at the restriction site and indicates no methylation, Type 2 is when MspI does cut and HpaII does not cut, indicating the restriction site has a methylated internal cytosine C; Type 3 is when MspI does not cut and HpaII does cut indicating the restriction site has hemi-methylation; and Type 4 is when neither enzyme cuts indicating either both cytosines are methylated or the restriction site has mutated (Richards, Schrey & Pigliucci, 2012) (conservatively, we treated Type 4 as missing data because the underlying methylation state cannot be determined; for example see Richards et al., 2012).
We performed MS-AFLP following the protocol used by Richards et al., (2012). For the MS-AFLP analysis, DNA was extracted using the Gentra Puregene tissue kit (Qiagen, Valencia, CA, USA) and was stored in 40 µl of TE buffer.  For both day 3 and day 11 samples, we digested approximately 250ng of genomic DNA at 37º C for 3 h in paired reactions: one with EcoRI and MspI, the other with EcoRI and HpaII. We immediately followed the restriction digest with adaptor ligation with EcoRI and MspI/HpaII adaptors at 16-20 h at 16º C (Supplementary Material Table 2, all primer and adapter sequences). After adaptor ligation, we conducted pre-selective PCR with EcoRI+1, MspI/HpaII+0 pre-selective primers (Supplementary Material Table 2) at the following PCR conditions: 75º C for 2 min; 20 cycles of 94º C for 30 s, 56º C for 30s, 75º C for 2 min, final extension at 60º C for 30 min and 4º C hold. Following pre-selective PCR, we conducted selective PCR by multiplexing 6-FAM fluorescently labelled EcoRI+AGC primers with HEX fluorescently labelled EcoRI+ACG primers and unlabeled primers HpaII/MspI+TCAT (Supplemental Table 2) at the following PCR conditions 94º C for 2 min, 8 cycles of 94º C 30 s, 65º C 30 s 72º C 2 min (dropping the annealing temperature 1º each cycle), 31 cycles of 94º C 30 s, 56º C 30 s 72º C 2 min, final extension of 60º C 5 min and a 4º C hold. We sent the selective PCR products to Macrogen Facilities (South Korea) for fragment analysis on an ABI 3130XL.
We used PEAKSCANNER v 1.0 (Applied Biosystems) to analyse resultant gel files and define fragment sizes, and RAWGENO (Arrigo et al., 2012) to define bands. We pooled data into two categories: methylated (Type 2 and Type 3) or not methylated (Type 1). Throughout, we refer to a MS-AFLP locus to indicate a particular sized band resolved in the selective PCR.
To ensure scores were consistent, we validated our MS-AFLP results by duplicating the entire protocol for 30 random individuals. We identified bands that consistently occurred, and we eliminated bands that were inconsistently amplified or occurred at highly variable intensities. We conducted all analyses using a binary haplotype-binding pattern (methylated 1, not methylated 0) for 92 verified consistent CpG sites between 50 and 500 base pairs for each nestling. We calculated percentage of genome wide DNAm as the proportion of the 92 loci that were methylated for each sample, we refer to this throughout as DNAm.