Machine learning for design of degenerate Cas13a crRNAs using lassa virus as a model of highly variable RNA target.

The design of minimum CRISPR RNA (crRNA) sets for detection of diverse RNA targets using sequence degeneracy has not been systematically addressed. We tested candidate degenerate Cas13a crRNA sets designed for detection of diverse RNA targets (Lassa virus). A decision tree machine learning (ML) algorithm (RuleFit) was applied to define the top attributes that determine the specificity of degenerate crRNAs to elicit collateral nuclease activity. Although the total number of mismatches (0-4) is important, the specificity depends as well on the spacing of mismatches, and their proximity to the 5' end of the spacer. We developed a predictive algorithm for design of candidate degenerate crRNA sets, allowing improved discrimination between "included" and "excluded" groups of related target sequences. A single degenerate crRNA set adhering to these rules detected representatives of all Lassa lineages. Our general ML approach may be applied to the design of degenerate crRNA sets for any CRISPR/Cas system.

Large scale screening of CRISPR guide RNAs using an optimized high throughput robotics system.

All CRISPR/CAS systems utilize CRISPR guide RNAs (crRNAs), the design of which depend on the type of CAS protein, genetic target and the environment/matrix. While machine learning approaches have recently been developed to optimize some crRNA designs, candidate crRNAs must still be screened for efficacy under relevant conditions. Here, we demonstrate a high-throughput method to screen hundreds of candidate crRNAs for activation of Cas13a collateral RNA cleavage. Entire regions of a model gene transcript (Y. pestis lcrV gene) were tiled to produce overlapping crRNA sets. We tested for possible effects that included crRNA/target sequence, size and secondary structures, and the commercial source of DNA oligomers used to generate crRNAs. Detection of a 981 nt target RNA was initially successful with 271 out of 296 tested guide RNAs, and that was improved to 287 out of 296 (97%) after protocol optimizations. For this specific example, we determined that crRNA efficacy did not strongly depend on the target region or crRNA physical properties, but was dependent on the source of DNA oligomers used for RNA preparation. Our high-throughput methods for screening crRNAs has general applicability to the optimization of Cas12 and Cas13 guide RNA designs.

Outbreak of mupirocin-resistant staphylococci in a hospital in Warsaw, Poland, due to plasmid transmission and clonal spread of several strains.

An outbreak of mupirocin-resistant (MuR) staphylococci was investigated in two wards of a large hospital in Warsaw, Poland. Fifty-three MuR isolates of Staphylococcus aureus, S. epidermidis, S. haemolyticus, S. xylosus, and S. capitis were identified over a 17-month survey which was carried out after introduction of the drug for the treatment of skin infections. The isolates were collected from patients with infections, environmental samples, and carriers; they constituted 19.5% of all staphylococcal isolates identified in the two wards during that time. Almost all the MuR isolates were also resistant to methicillin (methicillin-resistant S. aureus and methicillin-resistant coagulase-negative staphylococci). Seven of the outbreak isolates expressed a low-level-resistance phenotype (MuL), whereas the remaining majority of isolates were found to be highly resistant to mupirocin (MuH). The mupA gene, responsible for the MuH phenotype, has been assigned to three different polymorphic loci among the strains in the collection analyzed. The predominant polymorph, polymorph I (characterized by a mupA-containing EcoRI DNA fragment of about 16 kb), was located on a specific plasmid which was widely distributed among the entire staphylococcal population. All MuR S. aureus isolates were found to represent a single epidemic strain, which was clonally disseminated in both wards. The S. epidermidis population was much more diverse; however, at least four clusters of closely related isolates were identified, which suggested that some strains of this species were also clonally spread in the hospital environment. Six isolates of S. epidermidis were demonstrated to express the MuL and MuH resistance mechanisms simultaneously, and this is the first identification of such dual MuR phenotype-bearing strains. The outbreak was attributed to a high level and inappropriate use of mupirocin, and as a result the dermatological formulation of the drug has been removed from the hospital formulary.

Comparison of genetic characteristics of MRSA strains present in a Warsaw hospital in 1992 and 1996.

Nine isolates of methicillin-resistant Staphylococcus aureus (MRSA) collected in a Warsaw hospital in 1996 were typed by phenotypic (resistograms) and genotypic (PFGE and plasmid restriction analysis-REAP) methods. Twenty-four (MRSA) strains collected in this hospital during a period of the same duration in 1992 and typed earlier using resistograms and PFGE were also typed by REAP. Comparison of typing results obtained for isolates from 1992 and 1996 showed that strains characterised by PFGE patterns of two distinct types described as specific of the two clonally related groups of Polish MRSA in a multicentre study in 1992 are continuously present in the hospital. However, MRSA strains representing PFGE patterns not observed before were also found within the collection from 1996. REAP typing has proved to have a discriminatory power similar to that of PFGE analysis. Nevertheless, due to the lack of plasmids or difficulties in plasmid DNA isolation in 3 out of 33 studied strains, the typability of REAP turned out to be lower than that of PFGE.