Species-level salivary microbiome signatures in Sjögren's disease and the taxa obscured by short amplicons
Year: 2026
Type: software
Abstract: This deposit contains the analysis and figure code for a study of the salivary microbiota in Sjögren's disease. Unstimulated whole saliva from 124 women with the disease and 45 women without it was sequenced across the full 16S rRNA gene on PacBio Sequel IIe. Two criteria fixed in advance were intersected to define a panel of 22 species: a cascade of case-control filters covering abundance, prevalence, effect size and false discovery rate, and agreement of the species-level assignment across the eHOMD, SILVA and GTDB reference taxonomies. The code reproduces every quantity the paper reports. It covers the cohort freeze and its provenance record; alpha and beta diversity with PERMANOVA and a PERMDISP on the spatial median, both implemented here rather than called from a package; panel selection with label permutation of the whole cascade, split-half replication, a centred log-ratio check and a Dirichlet Monte Carlo analysis written from the published description; detection prevalence against a null that conditions on each library's own sequencing depth; correlation of a panel score with pre-specified clinical domains, with null results expressed as equivalence bounds and minimum detectable effects; and replication of the direction of genus-level effects in three published cohorts whose abundance tables were regenerated from the archived reads. It also contains the in-silico truncation experiment that gives the paper its second half. Reference sequences are cut with the same primer convention applied to the reads and compared with the sequences of every other named taxon, so that each panel species can be asked whether it stays distinguishable in V1-V2, V1-V3 and V3-V4. Identity is decided by global alignment, with the same method applied to every region so that a contrast between regions cannot be an artefact of measurement. The same procedure is then applied to every binomially named species in the database to give a background rate. Sixty-three analysis scripts and eighteen figure scripts are included, with a README mapping them to the sections of the paper. The raw sequences are in the NCBI Sequence Read Archive under BioProject accessions PRJNA1354591 and PRJNA1445968; the participant clinical record is not distributed.
Language: English
Subfield: Physiology (medical)
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7204253663