Figure 4. Characterization of U-DNA enrichment patterns.

(A) GIGGLE search was performed with interval (bed) files of uracil enriched regions on a set of HCT116 related ChIP-seq and DIP-seq experimental data (for details see Supplementary file 3). Factors corresponding to the top 10 hits for each sample were selected. GIGGLE scores between all seven samples and all experiments corresponding to these factors were plotted excluding those, where data were not informative (data are found in Supplementary file 3-table 1). Source data are available in Figure 4—source data 1. Histone marks and the transcription factors, SP1 and TCF7L2 are categorized depending on their occurrence in transcriptionally active or repressive regions. Notably, some of them have plastic behavior allowing either transcriptionally active or repressive function. U-DNA-Seq samples are as follows: non-treated wild type (WT, red), non-treated UGI-expressing (NT_UGI, orange), 5FdUR treated UGI-expressing (5FdUR_UGI, green) and RTX treated UGI-expressing (RTX_UGI, blue) HCT116 cells, and their MMR proficient counterparts (NT_UGI_MMR, yellow; 5FdUR_UGI_MMR, light green; RTX_UGI_MMR, light blue). GIGGLE scores are also indicated for our own H3K36me3 ChIP-seq experiments (RTX_UGI sample: empty squares, NT_UGI sample: empty triangles). The tendencies are even more pronounced if the RTX treated U-DNA-Seq is compared with the RTX treated ChIP-seq or if the non-treated U-DNA-Seq is compared with the non-treated ChIP-seq data. (B) Genome segmentation analysis was performed on signal tracks of 22 ChIP-seq data available for HCT116 cells in the ENCODE database, on our own ChIP-seq data for H3K36me3, and on the seven U-DNA enrichment profiles (bold). The Segway train was performed with 25 labels and the corresponding genomic segments were identified with Segway annotate (Chan et al., 2018). The signal distribution data were calculated using Segtools (Buske et al., 2011), and plotted using python seaborn/matplotlib modules (Hunter, 2007). Source data are available in Figure 4—source data 2. Details including the applied command lines are provided in Supplementary file 3. The color-code is applied for each factor (rows) independently, from the minimum to the maximum value as indicated. (C) Correlation with genomic features. Interval (bed) files of genomic features were obtained from UCSC, Ensembl, and ReplicationDomain databases (for details see Supplementary file 4-table 1), and correlation with interval files of uracil regions were analyzed using bedtools annotate software (details are provided in Supplementary file 4). Numbers of overlapping base pairs were summarized for each pair of interval files, and scores were calculated according the formula: (baseNo_overlap/baseNo_sample_file) * (baseNo_overlap/baseNo_feature_file) * 10000. Heatmap was created based on fold increase of the scores compared to the corresponding WT scores. Sizes of interval files in number of base pairs are also given in the second column and the second line. Upon drug treatments, a clear shift from non-coding/heterochromatic/late replicated segments towards more active/coding/euchromatic/early replicated segments can be seen. CDS, coding sequence; SINE, short interspersed element; LTR, long terminal repeat; LINE, long interspersed element; cytoBand, cytogenetic chromosome band negatively (gneg) or positively (gpos) stained by Giemsa; repl. timing, replication timing; DNaseHS, DNase hypersensitive site. (D) Correlation analysis with replication timing. Replication timing data (bigWig files with 5000 bp binsize) specific for HCT116 were downloaded from ReplicationDomain database (Weddington et al., 2008). Data bins were distributed to 10 equal size groups according to replication timing from early to late. Then log2 uracil enrichment signals for these data bin groups were plotted for each sample using R (Supplementary file 5). Source data are available in Figure 4—source data 3.

Figure 4—figure supplement 1. Comparison of our own H3K36me3 ChIP-seq data to each other and to the ENCODE data using Pearson correlation.

ChIP-seq experiments were performed for H3K36me3 in UGI-expressing non-treated (NT_UGI_H3K36me3) and RTX treated (RTX_UGI_H3K36me3) HCT116 cells (Materials and methods). Fold change over control tracks were calculated and compared to HCT116 specific H3K36me3 ChIP-seq data from the ENCODE database (ENCFF514ZYW (merged), ENCFF334KFI (rep1), ENCFF238GBP (rep2)) using Pearson correlation calculated by deepTools multiBigwigSummary and plotCorrelation (Ramírez et al., 2016) as described in Supplementary file 3. Notably, differences between our non-treated and RTX treated samples are not higher than difference between replicates of the ENCODE data.

Figure 4—figure supplement 2. IGV view of log2 ratio and regions of uracil enrichment on chromosome 1 (for all chromosomes see Supplementary file 2).

At the upper part cytogenetic bands by Giemsa staining are visible, the staining intensity (from white to black) correlates to the chromatin structure. Log2 ratio tracks (enriched coverage/input coverage, computed by deepTools/bigwigCompare) are colored by samples (non-treated wild type K562 cells (K562, brown), non-treated wild type (WT, red), non-treated UGI-expressing (NT_UGI, orange), 5FdUR treated UGI-expressing (5FdUR_UGI, green), RTX treated UGI-expressing (RTX_UGI, blue) HCT116 cells, and their MMR proficient variants (NT_UGI_MMR, yellow; 5FdUR_UGI_MMR, light green; RTX_UGI_MMR, light blue)). The negative values of log2 tracks are shown with a bit lighter color than the positives. For all the log2 ratio tracks, the same range was applied from −0.9 to 1. Regions of uracil enrichment derived from log2 ratio tracks are shown with the same color directly below the corresponding log2 tracks. The bottom track shows replication timing data (grey) for HCT116 downloaded from Replication Domain database (Weddington et al., 2008). Replication timing scores are derived from E/L Repli-seq experiments, where cycling cells are pulse-labeled with BrdU, then sorted to early and late S-phase fractions by flow cytometry, and BrdU labeled genomic DNA fragments are pulled down and subjected to NGS. Signal tracks are computed from the read coverages in early over late S-phase samples, therefore the higher score means earlier replication (Marchal et al., 2018). The scale goes from −2.5 to 5. Correlation among samples, replication timing and the cytogenetic bands are well visible (highlighted regions, Figure 4C–D, and Supplementary file 4).

Figure 4—figure supplement 3. Replication timing scores and AT content calculated on the genomic segments that were defined by the Segway analysis.

25 genomic segments from the Segway were analyzed. Average replication timing scores (gray bars) were calculated from two replicates available in Replication Domain database (Weddington et al., 2008), the standard deviation is also indicated. Replication timing scores are derived from E/L Repli-seq experiments, the higher score means earlier replication (Marchal et al., 2018). AT content of the segments (%, black square) were calculated using bedtools nuc (Quinlan and Hall, 2010) and the GRCh38 reference genome. Details of the calculations are provided in Supplementary file 4. For better comparison, signal distribution patterns of U-DNA-Seq samples derived from the Segway analysis (Figure 4B) are repeated at the bottom.