In the last reporting period, we have characterized how the chromatin structure ensembles we generate with a coarse-grained chromatin model and Monte Carlo simulations depend on the geometric parameters of chromatin, namely nucleosome wrapping by DNA and spacing of nucleosomes along the DNA. We found that the spacing and wrapping parameters can be determined from synthetic RICC-seq data predicted from a large number of simulated structures using a simple regression model. These parameter estimates help us match simulated structures against experimental data from different epigenetic states and can be used to build chromatin structure ensembles based on nucleosome spacing measurements from a variety of orthogonal epigenomic methods, such as MNase-seq.
Continuing to work with the version of RITRACKS described in the last reporting period, we have made an additional update to add elastic scattering cross-sections for DNA and estimate them for amino acids.
We have then run simulations with photons and several different ions of varying LET values to benchmark the code against other codes and against experimental data. These simulations were performed with single nucleosomes. Chromatin fiber simulations with multiple nucleosomes and different densities of nucleosomes will be started in the next month.
We have performed simulations with and without histones and observed that the yield of DNA breaks is approximately twice without histones as with histones, showing that our simulation recapitulates the well-known role of histones in protecting genomic DNA from radiation. Although the yield of double-strand breaks (DSBs) without histones is higher than predicted by other codes and observed in experiments, when histones are incorporated, the RITRACKS predictions for DSB yield agree with the other codes and experiments.
On the experimental side, we have adapted two protocols for use with cells irradiated with ionizing radiation: END-seq, which maps DNA DSBs, and GLOE-seq, which maps DNA single-strand breaks (SSBs), onto genomic DNA coordinates. These methods use single-ended mapping of breaks and do not depend on regions of high break density or on spatially correlated breaks to generate DNA fragments that can be sequenced, in contrast to RICC-seq, the method we were using in the prior reporting period.
In pilot experiments with X-rays, we have shown that by combining these protocols with normalization by irradiated genomic DNA controls, we can obtain estimates of DNA break density by epigenetic state that is different from that in scrambled genomic feature controls.
Our results from X-ray and preliminary ion experiments indicate that the density of both SSBs and DSBs varies between epigenetic states by approximately 10%, with less compact states, such as active promoters, exhibiting the highest break densities.
We have performed 200 Gy gamma ray and Fe ion irradiations at NASA Space Radiation Laboratory (NSRL) on four cell types: K562 leukemia cells, BJ fibroblasts, IMR90 fibroblasts, and RPE-1 retinal pigment epithelial cells. The DNA sequencing libraries resulting from these experiments are still being processed and we anticipate that sequencing data will be available for analysis in late spring 2024.
|