Microbial Genomes in Prostate Cancer Sequencing

Evidence

This single-study explainer covers a descriptive metagenomic analysis of non-human reads from whole-genome sequencing data associated with prostate cancer. The material came from blood, saliva, and prostate tumour tissue in the Pan-Prostate Cancer Group cohort, and the analysis included 247 samples. The study’s stated aim was characterization of microbial genome recovery and taxonomic diversity from sequencing data intended principally for human DNA analysis, rather than testing a diagnostic, prognostic, or therapeutic intervention. pmid:42901952

Investigators quality-trimmed unmapped reads from each included sample and assembled them with MetaSPAdes. Taxonomic assignment used GTDB-Tk. The investigators defined putative unidentified bacterial species when GTDB-Tk did not identify an existing organism at an average nucleotide identity of 95% or greater under default parameters. Across the analyzed samples, the median number of reads per sample was 4,952,049, with first and third quartiles of 2,949,908 and 9,987,724. These procedures describe genome reconstruction from residual non-human sequencing reads; they do not directly measure bacterial viability, tissue localization, or biological activity. pmid:42901952

The assembly process reconstructed 249 highly complete bacterial genomes. Recovery was reported predominantly from oral samples. The study also reported 18 bacterial species from the oral cavity without species-level classification, alongside a range of known species. This distribution is important for interpreting the dataset because the specimens were not interchangeable biological matrices: saliva, blood, and tumour tissue were all included, while the recovered genomes were concentrated in oral material. pmid:42901952

Saliva contained a median of 59,184,701 unmapped reads, with first and third quartiles of 35,283,741 and 107,635,982, compared with primary tumours, for which the reported comparison had a Wilcoxon rank-sum test P value of 4.58 × 10 −15. In contrast, the reported difference in unmapped reads between primary tumours and metastatic tumours was limited and not statistically significant, with Wilcoxon rank sum test P = 0.13. The reported read comparisons characterize the sequencing material available for assembly; they are not direct comparisons of microbial disease effects. pmid:42901952

Among the 249 recovered highly complete microbial genomes, 24/249 (9.6%) were not classified in existing species clusters and were considered unique by the study’s stated GTDB-Tk criterion. All of these genomes originated from saliva samples. Their estimated completeness had a median of 94.47%, with first and third quartiles of 92.27 and 99.23, and their median contamination estimate was 0.24%, with first and third quartiles of 0.03 and 0.69. The most commonly recovered unknown candidatus species belonged to the Campylobacter_B genus and represented one species recovered from five different patients. pmid:42901952

The taxonomic results also included Sphingomonas assembled in eight separate tumour samples. The authors noted that this genus is among clades frequently described as common sequencing contaminants, which places that tumour finding in a contamination-aware interpretive context. Separately, the study reported that a large majority of genomes with high completeness and low contamination were retrieved from saliva rather than tumour samples. The reported recovery pattern therefore distinguishes where assemblies were obtained within this sample set, not whether a recovered organism has a role in prostate cancer. pmid:42901952

Analysis — Interpretation of microbial genome recovery

The observed pattern is consistent with the study design and sample selection: microbial genome reconstruction was more productive in saliva than in tumour material, while tumour-associated assemblies were less common. The authors attributed the greater propensity for assembly in saliva to the oral cavity’s higher microbial biomass relative to tumours. That explanation addresses technical and biological context for recovery, but it does not turn sample-type differences in unmapped reads or assembly yield into evidence that particular bacteria contribute to prostate cancer. Likewise, identification of known organisms, unclassified oral taxa, and Sphingomonas in tumour samples is a taxonomic observation within the tested sequencing dataset. The analysis did not compare healthy and disease groups, and its stated future work included such comparisons. Accordingly, the results describe microbial genomic content recoverable from these samples and should be read as an observed pattern within a prostate-cancer-associated cohort, not as a causal account of cancer initiation or progression. pmid:42901952

Limitations

The descriptive metagenomic design cannot establish causation. It measured quality-trimmed unmapped reads and genome assemblies from blood, saliva, and prostate tumour whole-genome sequencing data; it did not measure a clinical outcome associated with the recovered bacterial genomes. The sample-source comparison also matters: saliva and primary tumours were compared for unmapped reads, and primary and metastatic tumours were compared separately, while healthy-versus-disease comparisons were identified by the authors as future work. These comparators constrain interpretation of differences across specimen types. pmid:42901952

Selection of specimen matrix can affect detectable microbial material. The authors stated that the oral cavity’s higher microbial biomass provides greater propensity to assemble microbial genomes from saliva than from tumours, which may contribute to the predominance of saliva-derived assemblies. In addition, the report characterized some taxa in relation to clades frequently described as sequencing contaminants and stated that future studies are required to ascertain the biological significance of the bacteria identified. These points leave the clinical and biological significance of the recovered genomes uncertain within the reported analysis. pmid:42901952