Alternative splicing (AS)
We used the JunctionSeq (Hartley and Mullikin, 2016)pipeline in order to detect genome-wide patterns of alternative spliced genes. AS is defined as the relative regulation of isoforms belonging to a multi-isoform gene with respect to a given biological condition (Hartley and Mullikin, 2015). The pipeline is based on differential usage calculated from both exon and junction feature coverages. The pipeline relies on the originally implemented method inDEXSeq (Anders et al., 2012), which tested differential usage of annotated exons, but extended to splice junction usage and both annotated and non-annotated splicing events. A new flattened GTF annotation file where overlapping features are not allowed was first generated using QoRTs (Hartley and Mullikin, 2015). All overlapping genes were merged as composed by a flat set of non-overlapping exons and splice junctions with unique identifiers.QoRTs was also used to generate a read count matrix for AS analysis, including three types of read counts per gene as estimated by exons, junction and gene level counts. The generated count matrix was then used by JunctionSeq R package (Hartley and Mullikin, 2016)to estimate differential exon and junction usage with respect to gene-wide expression. No read was counted more than once in the model since exon and junction dispersions are fitted independently. As for DE analyses, AS genes were detected if at least one exon or splice junction was differentially used as a result of parental or offspring acclimation at 36 °C vs 25 °C for TGP and WGP, respectively, using three biological replicates. Only features with p-values of < 0.05 after FDR correction were considered significant.