Background/Case Studies: Identifying genetic homology between bacterial strains isolated from blood products and recipients is essential for determining the causality of transfusion-transmitted bacterial infections (TTBI). Conventional methods, such as pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing (MLST), have limited discriminatory resolution owing to similar electrophoretic patterns among unrelated strains and frequent assignment of identical sequence types (STs) to genetically distinct strains. In this study, we performed whole-genome MLST (wgMLST) and average nucleotide identity (ANI) analyses to improve precision of strain-level homology analysis.
Study
Design/Methods: Three groups of bacterial species were analyzed: (a) isolates derived from platelet products and recipients of confirmed TTBI cases: Morganella morganii (n = 2), Streptococcus agalactiae, Streptococcus dysgalactiae, Serratia marcescens; (b) epidemiologically unrelated Staphylococcus aureus with identical ST (n = 15) ; and (c) S. aureus with different STs (n = 6). PFGE was performed by digesting whole-genomic DNA with restriction enzymes and comparing electrophoretic band patterns. MLST involved STs determination from sequences of six or seven housekeeping genes. For wgMLST, thousands of genes were aligned and compared following genome assembly. ANI analysis was conducted by dividing assembled genome sequences into 1,020 bp fragments and calculating the mean identity values across all fragments.
Results/Findings: PFGE and MLST failed to differentiate strains in groups (a) and (b). wgMLST detected 0 – 4 gene differences among group (a) strains, indicating a high degree of genetic similarity; however, 187 –287 gene differences were observed among group (b) and 1,776 – 1,787 among group (c). ANI analysis supported these findings: group (a) strains showed ≧ 99.9 % identity, group (b) ranged from 99.7– 99.8 %, and group (c) averaged 98.6 % identity. These results demonstrate that wgMLST and ANI are more effective than conventional methods in distinguishing genetically distinct strains. Conclusions: ANI values of ≧ 95 % generally indicate the same bacterial species, based on their equivalence to 70 % DNA–DNA hybridization. No standards exist for strain-level identity thresholds. Although only five bacterial species were analyzed, our findings suggest that strains with minimal wgMLST gene differences and ANI values ≧ 99.9 % are likely identical. These advanced methods offer superior discriminatory power over PFGE and MLST, and can play a crucial role in accurately resolving TTBI.
