A German team describes a web server for its DOGMA software, which is designed for quality analyses of transcriptomic and proteomic data. By comparing transcriptome or proteome data to a predetermined "core set" of conserved protein domains and domain arrangements gleaned from reference species, the researchers say, the DOGMA tool estimates the completeness and quality of the data on hand based on missing domains in the new data. DOGMA appeared to compare favorably to BUSCO when they applied both quality assessment tools to data from 153 eukaryotic species in the Ensemble genome browser. In particular, the authors suggest DOGMA may have an edge for looking at transcriptome or proteome quality in quickly-evolving species.
Researchers from Norway and the US present the latest version of a web tool for finding single guide RNA targets for CRISPR-based gene editing. The approach, known as CHOPCHOP version 3, builds on previous versions of the resource, the team says. It now includes tools for identifying potential Cas13 targets, for example, as well as improved methods for visualizing alternative isoforms and other factors that may CRISPR applications. "[T]he CRISPR-Cas system has been adapted for a wide selection of uses, and numerous factors influence each of these modes, necessitating the existence of intuitive software for target selection," the authors write. "This new update of the CHOPCHOP web tool incorporates new CRISPR-Cas targeting modes and predicts frameshift mutation frequency in an improved, user-friendly interface."
Finally, a team from Italy presents a web resource called Fido-SNP for predicting the consequences of coding and non-coding single nucleotide variants in the dog genome. The binary classifier approach behind Fido-SNP relies on a "gradient boosting," machine learning algorithm, the researchers explain, differentiating between pathogenic and benign SNPs based on training with almost 36,000 annotated human variants in ClinVar. Beyond the potential veterinary applications of the tool, the authors suggest that the approach used "can be seen as a proof of concept that the knowledge acquired through the high level of annotation of the human genome can be transferred and exploited to boost prediction performances in other species."