Curation involves three steps prior to final submission: identifying appropriate records, curating, and editing. This process is rigorously controlled at all steps to assure the accuracy of the information that is submitted into PhosphoSite^®. Each record must be separately curated and edited, providing multiple opportunities to verify the accuracy of the aggregated information.

Identifying appropriate records. Information is gathered from published literature and other sources. Published literature is searched semi-automatically with multiple intelligent search algorithms to identify reports that potentially identify phosphorylation sites in human, mouse or other species. Each identified report is then scanned by our highly trained curatorial staff (all with PhDs and extensive research experience in cell biology or related disciplines) to select only those papers that either identify new physiological phosphorylation sites or those that illuminate the biological function of the phosphorylation event. Records that are selected for inclusion into PhosphoSite are placed in the curatorial queue for processing. Note: while we gather records that describe both in vitro and in vivo phosphorylation events, we only finally submit records about in vitro sites when we have additional hard evidence that the site is also phosphorylated in vivo.

User participation. We strongly encourage our users to participate in this process. When you find that a particular in vivo site is missing from PhosphoSite, you are strongly urged to submit the relevant record to us. A feedback option is provided at the top of each page to make it easy for users to submit records that they feel should be curated into PhosphoSite. We will process all submitted information quickly. Users who submit more than 5 records that each contains one or more sites that are new to PhosphoSite will be publicly acknowledged as contributors if they wish.

Other sources of information may be proprietary and include the output from high-throughput phosphorylation site discovery programs at CST and other institutions. Initially this information will be shown to end users only when the site has been previously described in the public domain. Unrestricted access to this information may later become available on a subscription basis.

Curating. The first step in the process of curating an article is choosing a protein accession number from SwissProt or NCBI that will serve as the reference sequence for that protein in PhosphoSite. Information including the sequence, alternative names, and selected links is parsed into PhosphoSite. Similar information about orthologs and isoforms of the reference protein is also collected when necessary. Isoform information is curated only when the residue numbering used in the paper is isoform-specific and might otherwise be confusing to the end user.

After retrieving the protein information, the article is read to extract classes of information including:

1. the methodologies used to support the argument that the site is phosphorylated in vivo,


2. the putative kinase(s) or phosphatase(s) that may regulate the phosphorylation of the site, and


3. the downstream consequences of the phosphorylation event (activation, inhibition, degradation, etc.).

Since most types of experimental evidence about putative kinases/phosphatases and downstream effects are inferential, we also describe the methodologies that were used to support the author's conclusions. This information should help the end user evaluate the strength of the conclusions. For instance, if a paper infers that a particular phosphorylation reaction is regulated by protein kinase C and base this inference only on the treatment of cells with pharmacological activators such as phorbol esters, the inference may be considered weak at best. Note: the initial version of PhosphoSite will contain little of this inferential information; rather it focuses on identifying as many known in vivo sites as possible.

Editing. All curated records are reviewed for accuracy and consistency by an editor before final submission to PhosphoSite. The editor verifies all curated information from each paper and, after final submission, confirms that the phosphorylation sites (and their orthologous sites) are displayed properly to the user.

Apologies and Strategic Considerations. Initially, many phosphorylation sites will have only one journal article associated with them. This may not be the original article that proved that the site is phosphorylated in vivo, nor may it be an article that presents compelling arguments about the biological function of each site. As PhosphoSite matures, however, we intend to include the reports that originally demonstrated each phosphorylation event in vivo, and that illuminate the biological function of each site. This process will be considerably enhanced if knowledgeable users submit the relevant records to us. We will make every effort to quickly add these references into PhosphoSite.