Fig. 2

The pipeline. Sixteen transcriptomes were assembled using both de novo (Trinity de novo and SOAPdenovo–Trans) and reference based (Genome guided Trinity and Cufflinks) approaches, to reflect different infection levels (U1U2, L1L2, H1H2 and Master assembly: All U1U2L1L2H1H2). Only the four master assemblies were used in Differential Expression (DE) analysis, comparing the three treatments (U vs. L, U vs. H and L vs. H). Shared DE predictions of three DESeq2, EdgeR and EBSeq were identified for each transcriptome assembly. Before characterizing the DE genes using Gene ontology analysis, DE genes were pooled across assemblies and duplicates were removed. *: Trinity de novo master assembly was used for hierarchical clustering and qPCR validation