Researchers with BC Cancer and the University of British Columbia describe a pooled assembly-based algorithm called RNA-Bloom that's designed for putting together transcriptome assemblies from paired-end, short-read single-cell RNA sequence data or bulk RNA-seq data, regardless of whether a reference sequence is available. "Assembly with RNA-Bloom can be either reference-guided or reference-free, thus enabling unbiased discovery of novel isoforms or foreign transcripts," the authors write, noting that the algorithm outpaced several available reference-based and reference-free methods when it came to reconstructing isoforms from simulated or authentic single-cell RNA-seq data on thousands of mouse microglia cells or mouse embryonic stem cells.
A team from Sweden explores transcriptional features found in somatic motor neurons, which are susceptible to an autosomal recessive condition known as spinal muscular atrophy. With the help of laser capture microdissection and RNA-seq (LCM-Seq) and a gene correlation network analysis approach, the researchers examined gene expression over space and time in a range of neurons from mouse models of spinal muscular atrophy, uncovering apparent adaptations in ocular motor neurons that manage to dodge the effects of spinal muscular atrophy. "[T]his in-depth longitudinal transcriptomics analysis in [spinal muscular atrophy] reveals novel cell type-specific changes that, alone and combined, present compelling targets … for future gene therapy studies aimed toward preserving vulnerable motor neurons," they report.
Investigators in the US and UK describe graph-based software called Corticall for finding a range of de novo mutation types, from relatively simple to complex structural variants in repetitive regions, based on variant assembly from a combination of complete or draft reference sequences. "Our approach leverages long-haplotype data derived from any source (existing finished genomes, draft assemblies from third-generation sequencing, targeted sequencing of specific loci, etc.) to improve the assemblies of other short-read data sets," the team writes. The authors suggest that the Corticall method — which they used to profile de novo mutations in Plasmodium falciparum malaria parasites from a handful of crosses done in the lab — "opens the opportunity for multiple long-read data sets to be used to improve the connectivity of many more short-read assemblies.