O-GlcNAcAtlas tutorial

What is O-GlcNAcAtlas?

O-GlcNAcAtlas is a comprehensive, curated database of experimentally identified O-GlcNAc sites and proteins reported over the past several decades, beginning with the modification’s discovery in the early 1980s.

From search to a working set

Use Search a list of accessions on the search page to paste or upload up to 500 entries. Matching is exact and case-insensitive; isoforms stay separate. The batch report identifies duplicates, invalid inputs and accessions absent from this release. Comma, whitespace and semicolon delimiters are supported, with an optional accession column header.

Select proteins using the checkboxes or Select all matches. Selection persists across filters within that search. Save a named collection to revisit it later. My collections supports renaming, removing members and JSON backup/restore. Restoring a conflicting collection keeps both versions. Collections are stored only in this browser; clearing browser data removes them.

Compare two to four proteins to inspect their evidence, samples, sequence coverage and verified annotations. Each protein has its own coordinate axis. These tracks are not a sequence alignment, and the number of records is not a measurement of modification abundance. Exact coordinates and source papers are available in expandable lists.

Mapping review lists the source–sequence residue conflicts identified by the audit. Inspect the original peptide, its matching status and the source publication before recording a decision. A correction proposal requires a different positive coordinate, supporting PMID, reviewer and rationale. Decisions are local drafts, saved in an append-only history; they never modify published Atlas records. Export the history to share it with collaborators. Restoration requires the same audit revision and rejects altered historical entries.

Publication years come from verified cached PubMed metadata. Records with missing dates appear under Year unavailable. With both a PMID and year selected, the year must belong to that publication. Export matching evidence exports every original field for the sidebar-filtered evidence; each table's CSV additionally follows that table's own text filter.

Terms used in O-GlcNAcAtlas

O-GlcNAcAtlas provides the following information:

• Species. The species from which the original authors obtained samples used to identify O-GlcNAc sites, including human, mouse, rat, and others.

• Sample type. The tissue, cell, or other sample type used by the original authors.

• Accession. Protein accession identifier in UniProt or another database.

• Entry name. Protein entry name in UniProt or another database.

• Protein name. Protein name shown in UniProt or another database.

• Gene name. Gene name shown in UniProt or another database.

• Peptide sequence. Experimentally identified modified peptide sequence.

• Position in peptide. Modification position in the identified peptide sequence.

• Position in protein. Modification position in the protein sequence.

• Analytical throughput. LTP, low throughput; HTP, high throughput.

• Condition. Experimental comparison performed.

• log2 ratio. Log2 of the fold-change value.

• P value. Statistical value for a comparison.

• PMID. PubMed identifier for the original report that identified the O-GlcNAc sites or proteins.

• Comments. Curator comments for specific proteins, peptides, or sites.

How do I use O-GlcNAcAtlas?

1. Obtain a protein name, gene symbol, accession, or peptide sequence of interest.
2. Choose the matching search field. Specific identifiers such as a gene symbol or accession generally produce more focused results; for example, use MAPT or P10636 for human tau.
3. Start with the Proteins view for one entry per accession and species. Switch to Evidence records to see the individual source rows. Isoforms remain separate entries.
4. Use Refine results to filter by species, position annotation, experimental method, sample, ambiguity, publication or verified publication year. These filters apply to both views and are included in the page URL for sharing. The table search narrows the current view further.
5. Open an accession for its identity, sequence, peptides, evidence, and source publications. Numeric position counts and sequence highlights use positive integers; mixed-format source annotations remain available in the evidence.
6. Copy the visible rows or download the filtered result as CSV. On a record, Download complete record includes every source row and field, regardless of table filters or hidden columns. Reset table restores its original view.

Explore a protein modification map

The modification landscape places reported numeric positions on the available protein sequence. Circles indicate source records marked unambiguous; squares indicate records marked ambiguous. A position may appear on both tracks when its source records have different assignments. Counts describe records and distinct publications, not a confidence score.

1. Select a marker or use Reported position. The map opens the surrounding region, highlights the sequence context, and filters the peptide, experimental-evidence, and publication sections to that position. The arrow buttons move to the previous or next reported position.
2. Switch between Region and Full protein to change the scale. Show all positions restores every source record. Table filters can narrow the selected records further; Reset table clears only that table's controls.
3. Use Copy view link to share the current position. The page address records the selection, and browser Back and Forward restore earlier selections.
4. Use Download FASTA or Copy FASTA for the complete available sequence. Expand View complete protein sequence to see all reported numeric positions and the highlighted selection.

Records without a single numeric position, or whose positions fall outside the available sequence, are retained in the tables. Entries without an available sequence still provide their curated evidence. Download complete record and All record fields always retain every source row, including rows outside the selected position.

Take a protein set into analysis

Search suggestions show reported genes, protein names, accessions and species. Use the arrow keys and Enter to choose a suggestion, or submit a query in the selected search field. A gene name can match several proteins; check the accession and species before selecting.

1. Select proteins in search results and save them to a collection. Open My collections, select the proteins you want, and choose Analyze selected. A protein record also offers Analyze in Workbench.
2. Check the sequence-preparation outcomes. Compatible inputs from one species fill FASTA automatically. For mixed or incomplete sets, choose a human or mouse group. Missing sequences and unsupported species remain listed; isoforms are not substituted.
3. Inspect the FASTA and species model, then run the analysis. Matching Atlas sites link back to their experimental evidence. Predictions remain separate from evidence.

Browse the source studies

The publication directory searches the studies represented in Atlas by title, author, PMID or DOI. Each study page shows its reported proteins, numeric positions, methods and samples. You can save proteins to a collection or download all original study evidence. A source row citing several papers appears under each paper; study counts should not be added as independent observations.

Export figures and reproducible reports

Below a protein map, open Export a scientific figure. Set the first and last residue, use the current map region, or choose the full protein. Preview the figure and download SVG or a high-resolution PNG. The figure preserves the record-wide evidence filters, original coordinates, source-residue conflicts, compatible annotations and source information. Separate table searches do not change its evidence scope.

Export research report downloads a ZIP containing data, available sequences, filters, citations, versions and file checksums. Protein reports contain complete and record-wide-filtered evidence. Workbench reports contain the actual submitted FASTA, complete and filtered predictions, model configuration and source evidence from the completed run. Later edits to the input form do not change that run's report. Reports are created in your browser; keep a copy with your project files.

Search tips

• Use Search in the site header from any page, or press / outside an input field. Press Escape to close it.
• Check spelling carefully because there is no spell checker.
• Searches are not case-sensitive; MAPT and mapt are equivalent.

Follow evidence from a site to a study

On each protein record, the evidence filters combine reported method, sample type, site assignment, and publication with the selected position. Option counts reflect the other filters. A zero count means that combination has no matching source rows. Reported composite method names are retained as recorded.

The map, peptide table, evidence table, and publication list follow these filters. “Show all positions” clears only the position; “Reset view” clears the position and every evidence filter. The address and “Copy view link” preserve your selection, and browser Back restores it. Each table also has its own text search for inspecting that table.

“Download matching records” includes all original fields for the evidence filter selection. Table CSV buttons export the rows matching that table's current search. “Download complete record” always preserves all source rows and fields, regardless of filters.

Publication titles, authors, journal, date, and available DOI are cached from PubMed. “Copy citation” copies a plain-text citation; “View evidence” adds that PMID to the current filters and opens the evidence table. Missing metadata retains the original PMID link.

Interpret domains, regions, and sequence checks

The domains and regions panel uses cached UniProt annotations with exact coordinate boundaries. It includes domains, repeats, coiled coils, and annotated regions. The annotation menu and expandable list provide descriptions and coordinates without relying on the graphic. Source release, entry version, sequence version, and retrieval date accompany the panel.

Annotations are plotted only when the requested accession and the complete displayed sequence match the verified snapshot. A sequence change hides the coordinates. Canonical annotations are not transferred to a different isoform: an isoform is eligible only when UniProt explicitly identifies it as the displayed sequence and both sequences are identical.

A selected site may overlap an annotated region; that overlap does not establish a functional effect. Where a reported residue conflicts with the sequence, the panel does not assert a site-to-region association. Exact peptide searches can suggest candidate coordinates, but cannot independently verify a corrected biological site. All original evidence is preserved.

The complete mapping audit (JSON) records every conflicting position, per-record peptide checks, repeated or missing peptide matches, candidate coordinates, and unresolved cases.