New algorithm proves to be a powerful tool for detection and identification of novel bacterial organisms

By Pooja Toshniwal PahariaJan 10 2024Reviewed by Danielle Ellis, B.Sc.

In a recent study published in BMC Microbiology, researchers developed the Novel Organism Verification and Analysis (NOVA) algorithm to systematically analyze bacterial isolates unidentified using traditional methods such as matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) and partial 16S ribosomal ribonucleic acid (rRNA) gene sequencing using whole genome sequencing (WGS).

Background

Species identification is crucial in clinical bacteriology for treatment guidance. Conventional methods may not reliably identify some bacterial isolates due to a lack of reference data or uncharacterized organisms. Molecular techniques like 16S rRNA sequence analysis can reclassify and rename bacterial species. However, in a few situations, WGS is used due to higher species-level resolution.

About the study

In the present prospective study, researchers developed the NOVA algorithm to characterize strains unidentifiable by regular methods.

Researchers at the University Hospital Basel in Switzerland studied novel bacterial isolates, including clinically relevant ones and those whose identification was problematic. They provided genomic sequences of several bacterial organisms to increase epidemiological and taxonomic data.

The team also analyzed clinical data about patients and the clinical importance of the bacterial isolates to improve ecological and clinical knowledge of novel bacterial organisms. They conducted the study between December 2014 and January 2022 using clinical, molecular, and phenotypic information on bacterial species.

The team performed microscopy and bacterial cultures from various biological specimens followed by matrix-assisted laser desorption ionization-time of flight mass spectrometry to identify bacteria. They analyzed measurements using Bruker Daltonics data.

In the case of inability to reliably identify bacteria by mass spectrometry, divergent findings in the initial and subsequent hits, lack of validly published bacterial species, or no species-level identification, they subsequently analyzed the isolates using partial 16S rRNA polymerase chain reaction (PCR) followed by sequence analysis.

The team compared the resulting genomic sequences to those in the National Center for Biotechnology Information (NCBI) database. In the case of a minimum of seven gaps or mismatches (denoting ≤99% nucleotide similarity) in the sequences in comparison to the most similar correctly labeled species of bacteria, they included the isolates in the NOVA evaluation. They considered species validly documented in the List of Prokaryotic Names with Standing in Nomenclature (LPSN) German database as correctly described.

The team retrospectively extracted patient data from health records, and infectious disease specialists analyzed the microbiological reports with the clinical presentation of the patients. They estimated clinical relevance based on the clinical symptoms and signs, concomitant pathogen presence, bacterial pathogenicity, and clinical likelihood. 

Results

Sixty-one isolates were unidentifiable by standard methods, which were subsequently incorporated into the NOVA assessment. They identified 57% (n=35) of organisms as novel species of bacteria and 43% (n=26) of organisms denoted hard-to-identify isolates. Schaalia species and Corynebacterium species were the predominant genera. Twenty-seven of the 35 strains were isolated from deep tissue specimens or blood cultures, with seven out of 35 being clinically relevant.

The team identified four isolates as Gulosibacter hominis and one as Pseudoclavibacter triregionum. They identified two strains each within the genera Clostridium, Anaerococcus, Peptoniphilus, and Desulfovibrio, of which the two Corynebacterium pseudogenitalium isolates identified were published validly in recent times. The team detected one novel species of Citrobacter, Dermabacter, Lancefieldella, Helcococcus, Neisseria, Paenibacillus, Ochrobactrum, Porphyromonas, Pantoea, Pseudomonas, Pseudoclavibacter, Pusillimoas, Psychrobacter, Sneathia, Tessaracoccus, and Rothia.

In particular, the researchers identified one organism for the following bacterial species: Devosia equisanguinis, Cutibacterium modestum, Fenollaria massiliensis, Enterococcus dongliensis, Kingella pumchi, Kingella negevensis, Pantoea agglomerans, Mogibacter kristiansenii, Prevotella brunnea, Parvimonas parva, Pseudoramibacter alactolyticus, Pseudomonas yangonensis, Slackia exigua,  Vandamella animalimorsus, and Saezia sanguinis.

Medical histories and data on clinical importance were available in 47 out of 61 cases, with 15 out of 47 considering the respective bacteria clinically significant and 21 as not. The team classified three out of 15 cases as clinically significant with monomicrobial culture growth. Patient ages ranged from seven to 94 years, with 64% males and 36% females. Twenty-six isolates previously described but unidentifiable by standard techniques could be identified only by whole genome sequencing. These bacterial strains represented 19 validly published species and three yet to be validly published. Seventeen (65%) isolates showed Gram positivity, while nine (35%) were Gram-negative.

Conclusion

Overall, the study findings demonstrated the effectiveness of the new NOVA algorithm in detecting and identifying novel bacterial organisms that are difficult to characterize by routine diagnostic methods using WGS. The team identified 35 novel strains, seven clinically relevant, highlighting a wide range of undescribed pathogens yet to be defined. Corynebacterium was predominant among the 61 NOVA-classified isolates, with 11 being hard to detect and six representing novel species.