Baylor College of Medicine researchers introduce open-source software tools for teasing out cell-type specific methylation patterns based on bisulfite sequencing-based cytosine methylation data. While the "precise read-level imputation of methylation" (PReLIM) machine learning algorithm was developed to boost whole-genome bisulfite sequence coverage through individual read imputation, the team says, a related "cluster-based analysis of CpG methylation" (ClubBCpG) software offers a look at CpG methylation at the bisulfite sequence read level. "Our data indicate that, rather than stochastic variation, read-level CpG methylation patterns in tissue whole-genome bisulfite sequencing libraries reflect cell type," the authors report, noting that "these new computational tools should lead to an improved understanding of epigenetic regulation by DNA methylation."
A team from the Hebrew University of Jerusalem outlines a proteome-wide association study (PWAS) approach that they used to explore phenotype associations stemming from altered protein functions in simulated datasets or real datasets from the UK Biobank project. The PWAS method relies on a machine learning model for predicting genetic variant impacts on proteins, the researchers write, which can be applied to participants' genotyping profiles. The approach "is based on the premise that causal variants in coding regions affect phenotypes by altering the biochemical functions of the genes' protein products," they write, noting that "PWAS quantifies the extent to which proteins are damaged given an individuals' genotype."
Finally, a Yale University-led team shares an RNA sequencing pipeline for teasing apart host and microbial gene expression, and related host-microbe interactions, in non-invasively-collected sputum samples from individuals with or without asthma. By bringing in statistical modeling and strategies for reducing the "dimensionality" of the heterogeneous sputum RNA-seq data, the researchers developed a pipeline called LDA-link for interpreting host and microbe reads, which they validated using data from single-cell RNA-seq and imaging experiments. "By using the LDA-link, we identified associations that could be extracted from this heterogeneous dataset," the authors report, adding that "both fungi and bacteria showed these links, further highlighting the need to evaluate more than bacteria when performing microbiome experiments in the airway."