ProtAnnot reads an XML format defined in a ProtAnnot XML schema document that is available from the Genoviz SourceForge.net repository. To view the XSD document, visit the protannot resources directory and click the link labeled "view" to display the XML scheme document in your browser. The file contains numerous comments explaining the different data a ProtAnnot XML file presents.
We've developed some sample ProtAnnot files (see diff_motifs/hg19) representing gene structure and protein annotations from version 19 of the human genome (hg19).
To create these files, we obtained RefSeq gene annotations (in BED format) from UCSC Genome Informatics Table Browser, constructed a virtual transcript by concatening genomic exons, translated those virtual transcripts into protein, and then used the InterProScan Web service at EBI to annotate the proteins. We then combined all of these data into a single XML file (one per gene) for you to view in ProtAnnot.