maniFasta and the AllOralsDB: simplifying construction of comprehensive reference databases for metaproteomics

Metaproteomics aims to capture a taxonomically comprehensive snapshot of proteins in a sample. Design of reference databases is a key aspect of metaproteomic workflows, as these define what is ultimately seen. Databases tailored to focal biomes offer optimal performance, yet their construction often requires drawing on heterogeneous data sources, posing a challenge to reproducibility and documentation. Here we present maniFasta, a tool enabling users to generate standardized, reproducible, and robustly documented protein reference sets from diverse input sources and datatypes. Users provide information on their desired input types and sources, and the output is an integrated database comprising a protein sequence file (FASTA), with harmonized identifiers and standardized headers, and an associated provenance metadata table (manifest). We highlight the value of maniFasta in the context of salivary metaproteomics, addressing the need for a taxonomically comprehensive reference database. The AllOralsDB resource includes human proteins, as well as proteins from bacteria and archaea, fungi and other microeukaryotes, viruses and viroid-like elements, dietary sources, and common contaminants. Together, this work provides a community resource for oral and salivary metaproteomics (https://www.homd.org/ftp/AllOralsDB/), and a versatile and accessible tool for constructing protein databases for metaproteomics generally (https://github.com/KauffmanLab/maniFasta).