Fig. 6
RepEnTools screening revealed enrichment of hUHRF1 Tandem-Tudor Domain binding on SVAs and ERVs. A Domain structure of the UHRF1 protein containing a Ubiquitin-Like domain (UBL), a Tandem-Tudor domain (TTD), a Plant Homeodomain (PHD), a SET- and RING-associated (SRA), and a Really Interesting New Gene (RING) domain. Bottom: Scheme of the human TTD construct (Uniprot Q96T88, residues 126–280) used in a previous work with an N-terminal GST-tag in CIDOP-seq (Chromatin Interacting DOmain Precipitation) to selectively enrich HepG2 mononucleosomes [26]. B The volcano plot provided by RepEnTools reveals the reproducible enrichment and depletion of various repeat elements (REs) by analysis of TTD CIDOP signals versus input. The plot depicts significance log10(p) versus fold-change log2(ChIP/input) on the y-axis and the x-axis, respectively. The colour code indicates the False Discovery Rate (FDR) adjusted p-values ≤ 0.05 of enrichment and depletion with statistical significance, correcting for the multiple comparisons. REs with enrichment scores of log2 |(TTD/Input)|≤ 0.5 are not meaningful and are coloured grey. C The bar diagrams from the RepEnTools output reveal the repeat families most and least enriched in TTD CIDOP versus input. The top genome-wide hit, the SVA family, reveals enrichment in all SVA subfamilies, but strongest in SVA-F, the youngest and hominid-specific repeat [10]. The second most enriched family, the ERVK, reveals highest enrichment in the LTR22 subfamilies. ERVK depletions can be seen in Additional file 1: Fig. S8D. Of the ERV1 family, the highest enrichment is found on HERVE-internal regions. This also represents the most enriched individual RE genome-wide. In all bar diagrams, the bar indicate the mean, and the whiskers represent the standard deviation. All data with reproducibility p ≤ 0.05, n = 2 pairs of TTD/input. D Localised TTD enrichment is found at the 3′ end of SVA regions, confirming RepEnTools’ enrichment finding. Profile of all SVA models (pHMM) on chm13v2 (6274 regions), anchored to the 3′ end. RE track shows density of actual SVA annotated segments within the model. pHMM retrieved from RMSK bed12 output found on UCSC Table Browser [13, 29]. pHMM – profile Hidden Markov Model, RPKM—Reads per kilo base per million mapped reads. See also Additional file 1: Fig. S8E. For an illustration of the differences between pHMM of a RE and the actual RE see Additional file 1: Fig. S2A. E Broad TTD enrichment is centred on LTR22 regions, aka HERVK(HML-5) [41], confirming the enrichment findings by RepEnTools. Profile of all LTR22 models (781 regions). F Very strong TTD enrichment is seen at the center of HERVE-internal regions confirming the enrichment findings by RepEnTools. Profile of all HERVE-int models (142 regions). G TTD enrichment has a similar pattern at the center of the HERVE-int consensus (7.9 kb), when aligning the same datasets with bowtie2. This demonstrates that the findings of RepEnTools are reproducible even with a different aligner and reference. The peak in the middle of the HERVE pol gene overlaps the ORF of a 269 aa polypeptide bearing 92% similarity to a RNase H2-like domain found in reverse transcriptases. TTD and input data aligned by bowtie2 (fast, local), consensus sequence and gene positions retrieved from dfam [42]. See also Additional file 1: Fig. S8F, Additional file 2: Text S2. H The alignment strategy employed in RepEnTools (chm13v2—HISAT2) aligns a comparable number of reads to an alternative strategy (consensi—bowtie2), for a RE well represented by its consensus sequence (HERVE-int). This reproduction validates the strategies of RepEnTools, and demonstrates its advantages. I TTD CIDOP was reproduced with WT domain and the D142A binding deficient mutant. Assayed by qPCR, a H3K9me2 reporter locus and one for H3K4me3 demonstrated similar enrichments and depletions as shown for the samples that gave the HTS data [26]. All data from n = 2 biological replicates, bar indicates mean. J The new CIDOP-qPCR experiments corroborated the TTD enrichments and depletions reported by RepEnTools using carefully designed and validated qPCR assays on selected targets. CIDOP-qPCR was performed using TTD and chromatin from HepG2 cells in two biological replicates. Enrichment of TTD WT over D142A represents specific over unspecific pull-down. The assays shown here target HERVE-int, the most enriched RE genome-wide, and X4b, a member of TcMar, the most depleted family. For validation of the qPCR assays see Additional file 1: Fig. S9 and S10, and Additional file 3: Table S2. See also Additional file 1: Fig. S10E