Traceable Claims in Genomics

Interpretation & critique (visual-first, then evidence)

1. What a "clonal complex (CC)" is: A CC groups sequence types (STs) from MLST that share recent ancestry (eBURST clusters). CC labels (CC8, CC5, CC22, etc.) are central, stable population descriptors used in surveillance, outbreak tracing, resistance/virulence association, and comparative genomics Definition and role of CCs (JCM 2008).
2. Which CCs dominate globally (catalog evidence): Large-genome catalogs (98,950 GenBank assemblies) report CC8 (~20.6%), CC5 (~20.3%), CC22 (~8%); distribution strongly skewed and influenced by sampling bias (clinical/resourced-country genomes overrepresented) — interpret percentages as catalog composition, not unbiased global prevalence 98,950-genome CC breakdown (BMC Genomic Data 2025).
3. How to assign/validate CCs in practice: MLST (sequence of 7 housekeeping genes) -> ST, then eBURST groups STs to CCs. spa typing is portable and often reliably infers CCs (high concordance reported), and combining spa with PFGE gives maximal discrimination for local outbreak work; sequencing-based cgMLST provides higher resolution for WGS-era surveillance Typing-method recommendations (JCM 2008).
4. Limitations, biases, and blindspots:
   • Catalog sampling bias: GenBank-derived CC frequencies reflect where and what gets sequenced (clinical, outbreak, MDR strains), not random population prevalence — authors of the 98,950-genome resource explicitly warn about this Sampling bias and limitations (BMC Genomic Data 2025).
   • Concordance ≠ phylogeny: High concordance between spa and CC (eBURST) is useful but does not replace full phylogenomic analysis — recombination and convergent spa repeats can mislead.
   • Method trade-offs: PFGE remains highest-discrimination locally but is non-portable; spa & MLST/cgMLST are portable but lower resolution for outbreak fine-mapping unless combined with WGS-based approaches PFGE vs spa/MLST tradeoffs (JCM 2008).
5. Practical guidance (actionable):
   • If you need CC assignment for surveillance: use MLST/ST -> eBURST (or automated MLST/cgMLST pipelines) and cross-check with spa typing for portability; for historical PFGE datasets, combine PFGE+spa where possible to infer CCs (recommended by JCM 2008 authors).
   • For large-scale WGS surveillance: use cgMLST + MLST/CC annotation (as in the 98,950-genome database) but always account for metadata and sampling biases when estimating population prevalence.

Direct evidence used (key citations)

• Typing methods and CC inference (JCM 2008)
• 98,950-genome CC distribution (BMC Genomic Data 2025)
• PCR/ESI-MS CC assignment validation (JCM 2009)

Confidence, counterpoints & what would change this conclusion

Confidence that I (BGPT) am "familiar" with S. aureus clonal complexes: high (likelihood estimate reported below). Counterpoints that would materially change recommendations: discovery that spa repeat evolution systematically misassigns CCs across many lineages; a global, unbiased population sampling that shows a very different CC frequency distribution than GenBank; or method-specific systematic errors in MLST/cgMLST pipelines. Each claim above links to primary analyses and caveats cited.