A schematic representation of quantifying gene expression plasticity and genetic change in relevant genes. Experimental design, transcriptomic data of the flight and cardiac muscle were collected from tree sparrows obtained in their original lowland habitat (ancestral stage), experimental exposure to the hypoxic condition (plastic stage), and colonization to a high-elevation environment (colonized stage). Identify candidate genes, co-expressed genes that display correlation with hypoxia tolerance were identified by a WGCNA analysis and muscle phenotype associated genes were identified by correlating gene expression levels and each of the muscle phenotypes. Reinforcement/reversion plasticity, the genes with reinforcement or reversion plasticity were defined by comparing expression changes between the ancestral and plastic stages (plastic change), and between plastic and colonized stages (evolved change). The strength of reinforcement/reversion plasticity is measured by a range of thresholds, i.e., <100% and >50%, <150% and >100%, <200% and >150%, and >200%. Quantify genetic divergence, genetic divergence (i.e., FST) between lowland and highland populations was estimated for the genic region, 2kb up-stream and down-stream regions of the candidate genes. The empirical FST values were compared with permuted FST distributions generated from 100 random samplings from the genic background. A threshold of P<0.05 was used to determine statistical significance.

The WGCNA analysis identified co-expressed genes and reinforcement and reversion of expression plasticity. (a) Pearson linear correlation is used to identify genes with expression reinforcement and nonlinear regression to identify genes with expression reversion. Regulatory modules were identified as branches of the resulting cluster tree via the dynamic tree-cutting method and highly correlated modules (P<0.1) were merged. We used “GS > 1st quartile of GS” and “P<0.05” to identify important genes within modules (hub genes). (b) and (c) Six and nine modules respectively were identified to be associated with different stages for the flight and cardiac muscles (P<0.01). Within each module, colored dots show the genes with the expression levels significantly associated with the stages (gene significance function>1st quartile and P<0.05). (d) Frequencies of genes with reinforcement and reversion plasticity (>50%) and their subsets that acquire strong support in the parametric bootstrap analyses (≥ 950/1000). (e) and (f) Left, genes with expression plasticity being reinforced (red) or reversed (blue) at the colonized stage identified for the flight (e) and cardiac muscles (f), respectively. There are more genes showing reversion plasticity than those showing reinforcement plasticity. Two-tailed binomial test, NS, nonsignificant; ***, P < 0.001. Right, the FST values (vertical lines) significantly increase in the 2kb up-stream and/or down-stream regions of the genes having the magnitude of reinforcement/reversion plasticity > 200%. Vertical lines, the empirical FST values; shades, permutated FST distributions generated from 100 random samplings. NS, non-significant, *, P < 0.05; **, P < 0.01.

Genes with expression levels associated with muscle phenotypes and adaptive and maladaptive plasticity. (a) Gene expression profiles were correlated with muscle phenotypes. The direction of gene expression plasticity at the plastic stage matched the expectation for positive regulator (increased expression) and negative regulators (decreased expression) in the colonized tree sparrows was considered to indicate adaptive plasticity. Conversely, those that opposed this expectation were considered to be maladaptive, i.e., the direction of gene expression plasticity at the plastic stage showed decreased expression for the positive regulators and increased expression for negative regulators. A, ancestral stage; P, plastic stage; C, colonized stage. (b) and (c) Gene expression profiles were correlated with three flight muscle phenotypes (b, capillary per fiber, fiber area and perimeter) and two cardiac muscle phenotype (c, capillary density and fiber diameter). Red dots show genes with adaptive gene expression plasticity and blue dots show genes with maladaptive gene expression plasticity. (d) Frequencies of genes with adaptive and maladaptive plasticity (>50%) and their subsets that acquire strong support in the parametric bootstrap analyses (≥ 950/1000). (e) and (f) Left, candidate genes were classified to four categories with different magnitudes of adaptive or maladaptive plasticity. More genes show maladaptive plasticity than those show adaptive plasticity. Two-tailed binomial test, **, P < 0.01; ***, P < 0.001. Right, the empirical FST values significantly increased in the 2k up-stream and/or down-stream regions of the genes having adaptive/maladaptive plasticity <200% and >150% (cardiac muscle) and >200% (flight and cardiac muscles). (e) Flight muscle, (f) Cardiac muscle. Vertical lines, the empirical FST values; shades, permutated FST distributions generated from 100 times sampling. *, P < 0.05; **, P < 0.01.