Imputation-free reconstructions of three-dimensional chromosome architectures in human diploid single-cells using allele-specified contacts

Abstract

Single-cell Hi-C analysis of diploid human cells is difficult because of the lack of dense chromosome contact information and the presence of homologous chromosomes with very similar nucleotide sequences. Thus here, we propose a new algorithm to reconstruct the three-dimensional (3D) chromosomal architectures from the Hi-C dataset of single diploid human cells using allele-specific single-nucleotide variations (SNVs). We modified our recurrence plot-based algorithm, which is suitable for the estimation of the 3D chromosome structure from sparse Hi-C datasets, by newly incorporating a function of discriminating SNVs specific to each homologous chromosome. Here, we eventually regard a contact map as a recurrence plot. Importantly, the proposed method does not require any imputation for ambiguous segment information, but could efficiently reconstruct 3D chromosomal structures in single human diploid cells at a 1-Mb resolution. Datasets of segments without allele-specific SNVs, which were considered to be of little value, can also be used to validate the estimated chromosome structure. Introducing an additional mathematical measure called a refinement further improved the resolution to 40-kb or 100-kb. The reconstruction data supported the notion that human chromosomes form chromosomal territories and take fractal structures where the dimension for the underlying chromosome structure is a non-integer value.

Changes of chromosomal architecture before establishment of chromosome territories revealed by recurrence plot reconstruction

“…A summary of the RPR method is as follows (Fig. 1a) (2). First, we constructed a weighted network from a contact map obtained from a single diploid cell Hi-C dataset.…”

“…Tan et al has developed a modified scHi-C method for diploid cells called Dip-C that combines a transposon-based whole genome amplification procedure and an algorithm for imputation of the two chromosome haplotypes linked by each contact (35), although this algorithm contains some imputations that are not necessarily proven mathematically. We reported another algorithm, the Recurrence Plot-Based Reconstruction (RPR) method, that reconstructs the 3D chromosome structure from scHi-C data of either haploid (1) or diploid cells (2). This method is based on a recurrence plot (a tool of nonlinear time-series analysis for visualizing patterns within a time series) (1) and enables a unique 3D chromosome architecture to be reconstructed from even low-coverage DNA-contact information.…”

“…The RPR method has been mathematically proven (1, 37, 38). In contrast to Dip-C’s algorithm which assumes that paternal and maternal alleles are not spatially close to each other and do not have similar shapes, RPR algorithm only assumes that consecutive genomic regions are neighbors (1, 2).…”

“…However, reconstruction of nuclear architecture from single-cell Hi-C (scHi-C) data is challenging due to limited information of DNA contacts obtained from a single cell. We have previously developed the Recurrence Plot-Based Reconstruction (RPR) method for reconstructing three dimensional (3D) genomic structure from Hi-C data of single haploid cells (1) and diploid cells (2). This algorithm is based on a recurrence plot, a tool of nonlinear time-series analysis for visualizing patterns within a time series (3, 4), and enables the reconstruction of a unique 3D chromosome architecture even from low-coverage DNA contact information.…”

3D genome organization

“…In this 3D Genome Organization Collection, several groups have presented technical developments that increase the resolution power to interrogate the structure of the genome at a more detailed level. Hirata et al presented an imputation-free computational method for individually reconstructing the 3D architecture of each chromosome by integrating allele-specific single nucleotide variation data for each chromosome pair into a Hi-C dataset from single diploid cells 7 . Li et al introduced a chromatin packing domain structure, based on their observations employing chromatin scanning transmission electron microscopy (ChromSTEM), together with DNA-specific staining 8 .…”