Fig. 5

KLS is independent of canonical silencing pathways. a Bisulfite Sanger sequencing of nosP (301 bp on 3′-end of the insert). Methylation levels in all contexts significantly differed between active and silenced lines and also between individual lines (Additional file1: Table S5–6, Figure S3B-E). Error bars: Wilson 95% confidence intervals. b Pearson’s correlation analysis between 4C interaction frequencies with KEEs and pericentromeres (KEE-IF) and nosP methylation levels. Weak correlation was observed (red line). Non-significant correlation was observed when the highest methylated line (SG314) was omitted (blue line). c Correlation between kanamycin resistance phenotype and nosP methylation levels. Weak correlation was observed (red line). Non-significant correlation was observed when SG314 was omitted (blue line). d Differential analysis of sRNA-seq data. Genome-wide genomic bins (500 bp) exhibiting significant changes (logFC > 2; FDR < 0.01) between S- and A-lines (see also Additional file 1: Figure S4A). e Percentage of 21 nt and 24 nt sRNA-seq reads found within pROK2. For each genotype, biological triplicates were assessed (the number of reads were normalized by transgene copy number) (Additional file 1: Table S2). f Coverage of sRNA sequencing reads in 10 bp bins (21 nt–24 nt) mapping to the vector pROK2. Read numbers were normalized for pROK2 copy number. The reads of the three biological replicates were pooled. Right: Summary of first nucleotides of the reads. Asterisks mark significant deviations from equal distribution of nucleotides (chi-square test, P < 0.05)