The start page allows you to query the database. The menu on the left lets you access functions of the web site, regardless, if you are on the start page or not. The most important features that can be accessed from the menu are introduced in the following sections.

You can search for items in the database by entering your search term in the search bar (as shown in Figure 1 for gene CEBPB) and clicking the search icon on the right. Please note that your search term cannot be empty. This limit was imposed to ensure good search performance for a large quantity of users. Please note that by convention, we use genomic sequences, thus you must replace "U" in your nucleotide query by "T". By default, we ignore any special characters in your search for authors and publication titles: For example when searching for an author named "Muller", you will find "Muller" and "Müller". Searching for an author called "Müller" will find authors called "Müller" and "Muller".

Figure 1: The search interface.

Below the search bar, you can see grey panels; one for each view. We use the term view for a display of related items (publications, genes, transcripts, uORFs,...). Each view has a set of defined searchable fields. By de-(selecting) the checkboxes in front of the fields, you can refine your search results. By default, you are searching in all searchable fields from all views.
The panels will also provide you with the number of hits in the selected searchable fields from each view (see Figure 1). In our example, we searched for "CEBPB" which only gave us hits in the gene and publication views (see Figure 1). This is because, the search term only matched the fields in the gene and publication views. This does not mean that there are no entries in the other views that are related to CEBPB (e.g.: we have transcripts and uORFs for CEBPB).
For performance reasons, only the best 1000 hits per view are shown. By clicking the blue arrow button, you can directly jump to your search results. The arrow button only appears, if hits were found in the respective view.

If you choose to go from the search to the gene view, you will see those genes that matched your query. You can still navigate to any other view in the database by using the buttons at the top of the page. Buttons are greyed out, if there are no items in that particular view. All items that you will see in the other views are related to your hits in the gene view. Thus, navigating from your gene hits to the transcript view is not the same as performing the original search for CEBPB on the transcript view.

Figure 2: Controls in the gene view.

You can select one or multiple rows by ticking the checkboxes in front of the rows and clicking "Apply" in the top panel. This will remove any other rows from your current view. If you navigate to any other view, all items will be related to your selected rows. You also have designated buttons to tick all checkboxes on the current page ("Page+"), to uncheck all boxes ("None") or to uncheck all boxes on the current page ("Page-"). The number of checked boxes is shown next to these buttons. Selections can be undone using "Undo" or by clicking on one of your actions in the "Selection History". This will restore the results of that particular action.
On the top right, there are the pagination controls ("<<": first, "<": one page backward, ">": one page forward, ">>": last page), followed by the current page and total page count.
Some rows contain fields with many lines. We only display the first lines of data for each of these fields. Click on ▼ in the respective field to expand the row. Click on ▲ to shrink the fields of that row.
In the uORF and the variant view, we only display the first nucleotides of the sequences (indicated by "..."). You can click on the sequences to view the full sequences and to download them in FASTA format. The same applies to the amino acid sequences.
Links to external resources in the views are highlighted using .

Our uORF model allows for a quick analysis by eye. It shows the full TLS and CDS of transcripts with uORFs. ATG uORFs are highlighted in orange. Start codons that are shared with a CDS (see column description: "Shared start codon") are indicated by "*" behind the codon. You can access the model from the gene, transcript and uORF view (see Figure 3). Click on "Model" to see the results for all items in the current selection. However, there is a limit on how many items can be displayed at once. If you exceed the cutoff, you will get an error message and have to further limit your selection.
In the top menu, you can filter transcripts by NCBI ID and uORFs by start codon (the same principle as in the uORF view). By default, we don't include introns in the display and stop codons are hidden. You can change this behavior by selecting the respective checkboxes. To apply your selections, click "Filter".

Figure 3: The uORF model.

Each transcript has its own panel. Each panel shows the three reading frames of the transcript. Panels are independent. If you perform an action on one, the others will remain unaffected. If you move the cursor across the plot, the header of the respective panel will show you the current genomic position (0-based, half-open; see also column descriptions). To zoom in, click and hold the left mouse button and mark an area in the model: Mark a small area for a high magnification. Mark a large area for a more modest magnification of that area. Click "Reset" in the top right corner of the panel to reset the zoom. You can export your current view by using the "PNG" or "SVG" buttons in the top right corner of the panel.
If you want more information on a particular uORF, you can click on it to open an overlay window. To exit the model, click on any view button at the top of the page.

In the gene, transcript and uORF view, you can switch to the UCSC Genome Browser by clicking on the links in the gene symbol or the uORF ID columns. The link for gene symbol will only be present, if we analyzed the gene for uORFs. You will then see all transcripts and all uORFs for the associated gene. Nevertheless, if you entered the Genome Browser from the transcript or uORF view, the transcript or uORF that you clicked on will be highlighted. Using the Genome Browser, you can compare the uORFdb tracks to your own custom tracks or to tracks showing ribosome profiling data. A full list of features is beyond the scope of this tutorial, but the UCSC team has created a helpful manual.
If you want to continue working with the uORF sequences, you will find the track details page helpful. To access this page, click on one of our tracks in the Genome Browser. This will open a new page (see Figure 4) where we put direct links to the UCSC Table Browser. Using these links, you can easily create a custom download of the sequences or upload the sequences directly into the free GALAXY cloud service.

Figure 4: The track details page.

You can export all publications in your selection into a single RIS file by clicking the cite button in the publication view (there is an upper limit for the amount of citations that can be exported at once). RIS is a standard citation file format. Each record will be annotated with keywords from the database (KW in RIS). The keywords are a short summary of the publication by our uORF experts. They are extracted from the boolean (+/-) columns of the database (see also column descriptions below). If a "+" or a "-" is reported for a publication, the column name is used as a keyword for the citation file.
Our RIS file is fully compliant to its standard. Recognition of single fields, especially the keyword fields, depends on your reference manager. Nevertheless, we found that our citation files were processed as expected by Citavi, Endnote, and Zotero.

We publish the publications in uORFdb as an RSS feed that is updated once a day. RSS is a popular protocol to distribute new items to users (podcasts, new publications in journals, new publications matching a query in PubMed,...). All you need is a feed reader which is a special app or program that can understand RSS feeds. Feed readers are available for virtually any operating system, both on your local PC or on your smartphone/tablet (non-exhaustive list). Click the button "Publication Updates" in the left panel of our website. This will open a new tab. Don't worry about the content on this page. This is the RSS feed in a machine-readable format that is not necessarily good to read for humans. Just copy the URL of this page from the address bar of your browser. Now paste the URL into your feed reader. The way, how to do this, depends on your app. After that, you have successfully subscribed to our feed (see Figure 5, left). Every time we will add a new publication to the database, your app will notify you.
Some publication managers support RSS feeds which we feel is a particularly useful feature. The most mature implementation that we have seen so far is the one in Zotero. In Zotero, you can choose "New Library" --> "News Feed" --> "From URL" to open a dialog where you can paste the RSS URL.
From the feed view in Zotero, you can select a publication and inspect it. In the field for the abstract, there are the tags/keywords from uORFdb (see Export citations), followed by the actual abstract. From the top panel, you can chose to directly import the publication to your library (see Figure 5, right). For technical reasons, we cannot provide you the keywords from uORFdb, if you import an item from the RSS feed into your library. If you need the keywords, you can either add them manually or import the citation directly from the database, as explained in the previous section.

Figure 5: Left: A single publication from the RSS feed as displayed in the Feeder app on Android. Right: A single publication from the RSS feed as displayed in the Zotero reference manager.

Click on "Excel" or "CSV" in the export panel at the top of the uORF view, to export rows in your current selection in the respective format (there is an upper limit to the amount of uORFs that can be exported at once). This will export all columns, except for the links to dbSNP and ClinVar. The exonic nucleotide sequences will be the full sequences. The same applies to the amino acid sequences.

You can download all data that is stored in the database as TSV files. Click on "Downloads" in the main menu on the left of the web site. In the download tab, you will find further information on the type of content that can be downloaded and the file format (see also README file). One word of warning: Some of the exported files are huge, so please be aware that those should not be opened with a program like LibreOffice Calc or Excel. You should use the command line or specialized programs to analyze these files. If you want to explore the format of the downloads, we provide example files with only a few entries. These are generally safe to open with a graphical table viewer. All downloads can be verified using the MD5 checksum.

In the following you will find an in-depth description of all columns in the uORFdb web interface. The documentation is split in parts according to the views.

Please note the definitions of the most important technical terms:

If not indicated otherwise, the position names and the positions coordinates themselves are always based on the "+" strand. The smallest genomic coordinate is always the start, the largest genomic coordinate is the end. This intuitively makes sense on the "+" strand. On the "-" strand, however, it might be a little surprising. A uORF on the negative strand would end on its start position and start on its end position. If you want to have the "-" strand coordinates, please follow the steps in this resource.

By convention, we use genomic sequences and not mRNA sequences. This means that our sequences will not contain uracil, but thymine.

The results shown here are in whole or part based upon data generated by the TCGA Research Network.

Many of the following columns are not filled with text and/or numbers, but with plus and minus. A plus indicates positive evidence for a feature and a minus indicates negative evidence.

We regularly scan PubMed for the latest publications in the field. The version of the publication metadata (incl. author names) is the version that was available from PubMed at the time of the insert, but we will regularly scan PubMed for updated metadata.

We are using the NCBI taxonomy which we downloaded from the ftp server on the 06.05.2022.

The version of a gene for a publication is the current version of that gene on NCBI at time the publication was inserted.
For genes with transcripts and/or uORFs, the gene metadata is based on the All_Data.gene_info.gz and gene2refseq.gz files that we downloaded on the 04.05.2022.

uORFs predictions are based on the NM_* transcripts (in NCBI RefSeq Curated) and soft-masked genomes which we downloaded from UCSC on the 07.04.2022. We called uORFs for the following species and genome versions using custom scripts.

We then annotated uORFs with gene metadata using the files: All_Data.gene_info.gz and gene2refseq.gz; downloaded on the 04.05.2022. We filtered any transcripts and uORFs that belonged to pseudogenes or where the transcript accession was suppressed, withdrawn or renamed by NCBI. Also duplicate transcript IDs were removed (e.g. in the pseudoautosomal regions).

We analyzed BAM files from cancer patients provided by the TCGA Research Network. We analyzed the following cancer types: Breast Invasive Carcinoma (TCGA-BRCA), Colon Adenocarcinoma (TCGA-COAD), Acute Myeloid Leukemia (TCGA-LAML), Lung Adenocarcinoma (TCGA-LUAD), Prostate Adenocarcinoma (TCGA-PRAD), and Skin Cutaneous Melanoma (TCGA-SKCM). WGS BAM files were downloaded from the GDC Legacy Archive between the 28.05.2021 and the 09.01.2022. We realigned the BAM files to GRCh38.p13 (downloaded 15.06.2021). Realignment and subsequent quality control were performed according to a custom workflow heavily based on the GDC DNA-Seq Analysis Pipeline. 677 patients remained and entered the subsequent variant calling. Variants were called using Mutect2 (GATK v4.1.4.1, Java v1.10.11) with a custom pipeline based on the GATK best practices.
PASS Variants were left-aligned, normalized, and annotated using custom scripts and BCFtools v1.11. The annotation included dbSNP and ClinVar identifiers and allele frequencies from ExAC, gnomAD Genomes, and TopMed version 2 or 3. The metadata was based on dbSNP (downloaded 13.08.2021) and ClinVar VCFs (downloaded 13.01.2022), as well as dbSNP JSON files (downloaded 23.02.2022). In the functional prediction, we chose to include only variants where alternate and reference allele had the same length. This avoided frame shifts and issues with splicing sites. Prediction of the variant effect on the uORFs was performed using custom scripts.

Manske F, Ogoniak L, Jürgens L, Grundmann N, Makałowski W, Wethmar K. The new uORFdb: integrating literature, sequence, and variation data in a central hub for uORF research. Nucleic Acids Res. 2023 Jan 6;51(D1):D328-D336. doi: 10.1093/nar/gkac899.