Fig. 5

Cisplatin effects on tRNAs. A Total normalised reads that mapped to tRNAs in each condition: control condition (blue), cisplatin low-dose condition (green) and cisplatin high-dose condition (red). Data from generations were used as technical replicates and the two lineages per condition as biological replicates, i.e. for control: N = 21, for low dose: N = 20 and for high dose: N = 21. B The total normalised reads represented according to their mapping positions on the respective tRNA sequence and in each different condition: control condition (blue), cisplatin low-dose condition (green) and cisplatin high-dose condition (red). Data from generations were used as technical replicates and the two lineages per condition as biological replicates, i.e. for control: N = 21, for low dose: N = 20 and for high dose: N = 21. C Violin plot of the total normalised tRNAs mapping specifically to the 3′ half of the tRNAs in the different conditions: control condition (blue), cisplatin low-dose condition (green) and cisplatin high-dose condition (red). Data from generations were used as technical replicates and the two lineages per condition as biological replicates, i.e. for control: N = 21, for low dose: N = 20 and for high dose: N = 21. A significant increase was observed in both low- and high-dose cisplatin, with a trend towards increased reads in high-dose cisplatin (p < 2e − 16, Jonckheere test for ordered medians). D Bar plot showing the number of read counts associated to each kind of tRNAs. Data from all generations and all lineages were used giving 62 replicates. The tRNAs with the most associated fragments were GlycineGCC, GlutamineTTG and Glutamic acidCTC. E Volcano plot of the expression change between high dose and control (in log2(fold change)) for the different kinds of tRNAs. Data from generations were used as technical replicates and the two lineages per condition as biological replicates, i.e. for control: N = 21 and for high dose: N = 21. Points above the p < 0.05 line show a significant expression change for these corresponding tRNAs. Adjusted p-values were calculated using DESeq2. F Bar plot of the total normalised 3′ halves tRNAs reads and their association to each AGO protein. Each replicate per AGO was used. Both Ergo1 and Wago10 showed a high enrichment in tRNA 3′ halves relative to control Ips. G Boxplot of the number of new epimutations affecting tRNA fragments arising at each generation of the MA lines compared to the pre-mutation generation F0 and for each condition: control (blue), cisplatin low dose (green) and cisplatin high dose (red). Data from generations were used as technical replicates and the two lineages per condition as biological replicates, i.e. for control: N = 19, for low dose: N = 18, and for high dose: N = 19. A significant increase was observed in cisplatin high-dose condition (Kruskal–Wallis rank sum test followed by pairwise Wilcoxon test with Bonferroni correction: HD vs. C: p-value = 3.6e − 03, HD vs. LD: p-value = 0.02). H Survival curves representing the duration of epimutations in 22G-RNAs (red, N = 6) and 3′ halves tRNA fragments (blue, N = 6). No significant difference was observed between the two types of snRNAs (log-rank test, p-value = 0.09). Supporting data for the whole figure are available in the Excel file: “Additional file 23”