A custom microarray was designed from the sequence of the junco transcriptome. For 33,545 contigs, three unique probes are present on the array, while another 61 contigs are represented by two probes and 65 contigs are represented by one probe, accounting for 100,822 probes on the array. An additional 34,365 probes were selected from the remaining singletons (one probe per chosen singleton). The array also contains control probes and 2,604 random probes designed to reflect the genome nucleotide composition by Markov modeling to experimentally determine the appropriate thresholds that measure significant hybridization signals over the background. Thus, each sub-array consists of over 137,000 long-oligonucleotide (60 bp) probes, and 12 such sub-arrays are placed on each glass slide (Roche NimbleGen Inc., Madison, WI). The microarray platform is deposited at NCBI Gene Expression Omnibus (GEO; accession number GPL14995).
We collected adult dark-eyed juncos from breeding grounds near Mountain Lake Biological Station (Pembroke, VA) in mist-nets between May 7 and 14, 2010 and held them individually in a semi-naturalistic outdoor aviary where they were neither acoustically nor visually isolated from other juncos, as part of a larger experiment. On June 9 and 10 individuals were euthanized by overdose of isoflurane. Tissues, including whole brains, were collected rapidly and stored on powdered dry ice within 20 min post-mortem to ensure negligible RNA degradation [62 ]. Brains were dissected into 14 distinct regions using anatomical landmarks, following previously established methods [63 (link)] based on the zebra finch brain atlas. These brain regions included the hypothalamus and the ventral medial telencephalon (VmT), which primarily consists of the nucleus taeniae, the avian homologue of the medial amygdala [64 -66 (link)].
RNA from VmT, hypothalamus, liver, and pectoralis was extracted in TRIzol® following manufacturer's directions (Invitrogen, Carlsbad, CA). The microarray protocol follows previously published methods [67 ]. Briefly, total RNA was reverse-transcribed to ss-cDNA in the presence of oligodT primer and SuperScript II reverse transcriptase. This ss-cDNA was then converted to ds-RNA and labeled using CY-labeled random nonmer primer (either Cy3 or Cy5) and Klenow fragment (following NimbleGen labeling protocols). We then hybridized 4 g of each of two labeled samples (one Cy3, one Cy5) to each sub-array and followed manufacturer's directions for post-hybridization washing and scanning (Roche NimbleGen, Inc., Madison, WI). Imaging was accomplished by Axon GenePix 4200A scanner (Molecular Devices, Sunnyvale CA) with GenePix 6.0 software and data were extracted with NimbleScan 2.4 (Roche NimbleGen, Inc., Madison WI). Raw microarray data were processed with the limma package [68 ] in R version 2.9.0 [69 ] to normalize expression scores.
To determine if a gene was expressed, we calculated the 97.5% quantile for expression score of random probes in each individual as the cutoff for calling expression. Thus, for each individual, a called expression is significant at a p-value of 0.025. For each contig, we tested the median probe value against this threshold, and for singletons we used the single expression value. Because our design employed biological replicates, we called a contig or singleton expressed only if at least three of the six individuals in a group were called as expressed, thus reducing the p-value further to 0.0006 (the probability of obtaining at least three of six individuals called for expression of a random probe). From this, we determined whether or not a gene had expression support in any of our tissues-sex pairings, and which genes were restricted to expression in one sex.
We collected adult dark-eyed juncos from breeding grounds near Mountain Lake Biological Station (Pembroke, VA) in mist-nets between May 7 and 14, 2010 and held them individually in a semi-naturalistic outdoor aviary where they were neither acoustically nor visually isolated from other juncos, as part of a larger experiment. On June 9 and 10 individuals were euthanized by overdose of isoflurane. Tissues, including whole brains, were collected rapidly and stored on powdered dry ice within 20 min post-mortem to ensure negligible RNA degradation [62 ]. Brains were dissected into 14 distinct regions using anatomical landmarks, following previously established methods [63 (link)] based on the zebra finch brain atlas. These brain regions included the hypothalamus and the ventral medial telencephalon (VmT), which primarily consists of the nucleus taeniae, the avian homologue of the medial amygdala [64 -66 (link)].
RNA from VmT, hypothalamus, liver, and pectoralis was extracted in TRIzol® following manufacturer's directions (Invitrogen, Carlsbad, CA). The microarray protocol follows previously published methods [67 ]. Briefly, total RNA was reverse-transcribed to ss-cDNA in the presence of oligodT primer and SuperScript II reverse transcriptase. This ss-cDNA was then converted to ds-RNA and labeled using CY-labeled random nonmer primer (either Cy3 or Cy5) and Klenow fragment (following NimbleGen labeling protocols). We then hybridized 4 g of each of two labeled samples (one Cy3, one Cy5) to each sub-array and followed manufacturer's directions for post-hybridization washing and scanning (Roche NimbleGen, Inc., Madison, WI). Imaging was accomplished by Axon GenePix 4200A scanner (Molecular Devices, Sunnyvale CA) with GenePix 6.0 software and data were extracted with NimbleScan 2.4 (Roche NimbleGen, Inc., Madison WI). Raw microarray data were processed with the limma package [68 ] in R version 2.9.0 [69 ] to normalize expression scores.
To determine if a gene was expressed, we calculated the 97.5% quantile for expression score of random probes in each individual as the cutoff for calling expression. Thus, for each individual, a called expression is significant at a p-value of 0.025. For each contig, we tested the median probe value against this threshold, and for singletons we used the single expression value. Because our design employed biological replicates, we called a contig or singleton expressed only if at least three of the six individuals in a group were called as expressed, thus reducing the p-value further to 0.0006 (the probability of obtaining at least three of six individuals called for expression of a random probe). From this, we determined whether or not a gene had expression support in any of our tissues-sex pairings, and which genes were restricted to expression in one sex.
Free full text: Click here
