3.1 A comprehensive map of chromosomal contacts in chicken liver cells nuclei
To study the dynamic chromatin interactions in primary chicken liver cells from WCC and LDC, we carried out Hi-C experiments and sequenced approximately a total of 295 Gb high quality raw data. Alignment of the obtained sequences to the chicken reference genome (ftp://ftp.ensembl.org/pub/release-90/fasta/gallus_gallus/dna/) resulted in an average of ~263 million paired-end reads mapped for each individual, among which an average of ~207 million are intra-chromosomal reads(207,914,357 in LDC and 206,471,795 in WCC) (Table S1 and S2). Data processing also showed high quality of the Hi-C data, proving the successful performance of Hi-C experiments.
3.2 Identification and characterization ofcompartments in chicken liver cells
After data analysis, we acquired chromatin interaction heatmaps at 40-kb resolution for each sample (Fig. 1 and Fig. S1). The heatmaps displayed a typical plaid-pattern, which was demonstrated previously in chicken erythrocytes and fibroblasts Hi-C data (Veniamin et al., 2018), as well as in mammalian Hi-C data (Battulin et al., 2015). The plaid-pattern intimates the existence of large spatial compartments, such as compartments A/B. We know the genome consists of actively transcribed compartments A and inactive compartments B (Lieberman-Aiden et al., 2009), and switching of the compartments A/B are concerned with comparative changes of gene expression (Barutcu et al., 2015). We then determined the compartment types of the genome at 1 Mb resolution in liver cells, and found most genomic regions reserved the same compartments in LDC liver cells compared to WCC liver cells (Fig. 2 and Fig. 3). A total of 5% genomic regions switched between compartment A and compartment B, and as expected, we found more genes in the compartment A and they displayed higher transcriptional levels than those in compartment B (Fig. 4 and Fig. S2).
We next compared TADs in LDC and WCC liver cells. We called TADs from Hi-C interaction matrices at 40 kb resolution and identified 396, 386 TADs in LDC liver cells and WCC liver cells (Fig. 5 and Table S3), respectively, and with their median TAD sizes of 2.3 Mb (Fig. S3). Interestingly, previous studies have found TAD boundaries display an obvious enrichment of active genes when compared with randomly sampled genomic regions, which suggested a relationship between TAD formation and gene transcription (Wang et al., 2018). We thus investigated TAD boundary distribution, and 280, 261 TAD boundaries were identified in LDC and WCC genome, respectively (Table S4), with enriched transcription start sites in TAD boundary (Fig. S4), which further confirmed the reliability of the sequencing results.