Genomic and gene expression data

All Zymoseptoria spp. isolates used in this study were obtained from publicly available long-read sequencing-based genome assemblies. In total, we included 26 high-quality assemblies of Zymoseptoria spp. genomes from PacBio sequencing (Badet et al. 2020; Feurtey et al. 2020; Möller et al. 2020). These include the genomes of Z. ardabiliae Za17, Z. brevis Zb87; Z. passerinii Zpa63, and Z. pseudotritici Zp13 ( et al. 2020). We gathered a total of 22 Z. tritici genomes from strains isolated all over the world (Figure 1A) including eight isolates from Europe (the reference isolate IPO323 from the Netherlands); the four isolates Zt1A5, Zt1E4, Zt3D1, Zt3D7 (Plissonneau et al. 2018) from Switzerland, Zt05 from Denmark (Grandaubert et al. 2015), CRI10 from Czech Republic and UR95 from Ukraine (Badet et al. 2020); five isolates from America (CNR93 from Canada; I93, OregS90 from the USA; CH95 from Chile and Arg00 from Argentina) (Badet et al. 2020); one isolate from Oceania (Aus01 from Australia) (Badet et al. 2020); four isolates from Africa and the Middle-East (ISY92 from Israel; TN09 from Tunisia; KE94 from Kenya and YEQ92 from Yemen) (Badet et al. 2020); four isolates sampled in Iran (Zt10; Zt289; IR01_48b and Zt469) (Badet et al. 2020; Möller et al. 2020).

Transposable element content and identity variation in the Zymoseptoria genus. (A) Summary of the 26 long-read genome assemblies obtained from a worldwide collection of Zymoseptoria isolates, including the five species Z. ardabiliae, Z. brevis, Z. pseudotritici, Z. passerinii, and Z. tritici. Genome sizes are given in Mb. (B) Bars represent TE content (%) per genome estimated after REPET (Flutre et al. 2011) annotation. Colors represent TE order coverage with retrotransposons (LTR, LINE, and other class I orders in warm colors) and DNA transposons (TIR, MITE, and other class II orders in cold colors). Sequence identity distribution between TE copies and the respective consensus sequences (C) per genome and (D) per species and geographical origin. Each dot represents the median sequence identity of TE cluster. Boxplots are colored in regards to the species and isolate geographical origin.

Gene expression during wheat infection (Haueisen et al. 2019), updated expression profiles on the last versions of genome assemblies and new gene predictions of the three Z. tritici isolates IPO323, Zt05, and Zt10, were used as previously described (Feurtey et al. 2020). In summary, RNAseq was performed during wheat infection time-course using strand-specific RNA-libraries from Illumina HISeq2500 sequencing, with 100 pb single-end reads. A total of 89.5 to 147.5 million reads per sample were obtained (Haueisen et al. 2019). To simplify the expression data, we combined all time-points from wheat infection and calculated expression levels in transcripts per million (TPM) as described earlier (Feurtey et al. 2020).

Prediction of the “secretome” and effector repertoire were performed as follows: (1) secreted proteins were predicted using SignalP4.1 with “sensitive” parameter settings (Petersen et al. 2011), (2) predicted secreted proteins were extracted from the proteome and effectors were predicted using EffectorP2.0 (Sperschneider et al. 2018) for each of the 26 Zymoseptoria genomes.