An data/ignore.idx file specifies entries that are to be
    ignored and not processed. This is used to exclude incorrect entries to
    prevent processing failures. An entry made before the processing step
    prevents any processing of the entry. Entries can be added after the
    processing step has been done to prevent alignment or database loading.
    
    Ignore index entries are deleted from the database during load before
    any other processing occurs if they occur in the seq table.
    This allows the ignore mechanism to cleanup many difficult cases, such
    as a GenBank record correcting the type (mRNA <-> EST) of an entry.
    
The file is not associated with any particular release, so that entries apply across upgrades. The tab-seperated columns are:
    acc moddate srcdb comment
    acc - GenBank or RefSeq accessionmoddate - Modification date, in 2002-22-08 formatsrcdb - GenBank or RefSeqcomment - Some comment describing the problem.