National Institutes of Health researchers propose a quantitative MNase-seq approach that takes advantage of digestion features of a micrococcal nuclease (MNase) enzyme used to map nucleosomes. Because MNase enzyme has distinct effects at nucleosomes compared to linker regions, the team reasoned that it could get a glimpse at nucleosome occupancy across the genome by tracking nucleosome release over time in cells treated with the enzyme — a strategy it demonstrated in Drosophila fruit fly cells. "Our findings support the conclusions that have been drawn from nucleosome resolution imaging," the authors report, "in which the difference between heterochromatin and euchromatin is in the density of chromatin, but not in higher resolution features, such as the average diameter of the chromatin fiber."
A team from Leiden University Medical Center and Delft University of Technology in the Netherlands shares results from an evaluation of almost two dozen single-cell transcriptomics-based methods for automatically identifying and classifying cell identification approaches. Using a benchmarking workflow they call "Snakemake," and 27 publicly available single-cell RNA sequencing datasets generated for a range of mouse and human cell types and samples sizes using several scRNA-seq protocols, the researchers compared the performance of 22 supervised cell type classification approaches within individual datasets and across multiple datasets. "We find that most classifiers perform well on a variety of datasets," they report, "with decreased accuracy for complex datasets with overlapping classes or deep annotations.
University of California, Davis, researchers outline a computational strategy to tackle technical noise in single-cell sequencing data, including data generated by scRNA-seq and scATAC-seq. The "single-cell binary factor analysis," or scBFA, framework hinges on the notion that for "scRNA-seq datasets exhibiting high technical noise, dimensionality reduction using only the gene detection measurements is superior to the existing state-of-the-art methods that use both detection and quantification measurements," the team says. Using simulated and real datasets, the authors demonstrate their gene detection-based scBFA approach can identify cell type markers, follow apparent cell trajectories, and assess chromatin accessibility in samples with relatively high "quantification noise."