Search Failure Modes

1. Problem: No sequence match or identifier match at all.
   Possible explanation: You may have a new sequence that was not yet incorporated into ProPhylER, or your sequence is a predicted protein from a genome that is not part of ProPhylER. If you searched by PDB ID, the sequence in the PDB file does not have a match in ProPhylER's sequences.
   Possible solution: Attempt to find a related sequence by searching other databases and repeat the ProPhylER search with that related sequence.

2. Problem: ProPhylER BLAST found a match but the sequence is different.
   Possible explanations: ProPhylER has a closely related sequence from a different organism, or a closely related paralog, but not your sequence; you searched with a splice isoform that is not present in ProPhylER; or there are versioning/editing differences between ProPhylER's and your sequence.
   Possible solutions: See if the match ProPhylER gives you is nonetheless useful to you, or if you can translate the coordiates of the ProPhylER sequence/alignment to those of your sequence.

3. Problem: Your Uniprot or Ensembl Identifier Search found a match but ProPhylER uses a protein sequence that is different from what you expected.
   Possible explanation: ProPhylER unfortunately cannot yet handle different splice isoforms and, for every given gene, chooses one on the basis of a number of criteria (such as whether a Swissprot sequence exists for that particular splice form).
   Possible solution: See if the splice isoform ProPhylER uses can meet your need.

4. Problem: Your protein was found but ProPhylER does not have an analysis for it.
   Possible explanation: Your sequence may be part of ProPhylER, but it didn't make it into a final alignment. To understand why that may happen, please see Figure 1 below.
   Possible solution: Your sequence is a member of the cluster but was perhaps compromised in some way, leading to its elimination from the final alignment. But you may find a rather closely related sequence that is part of the alignment, which could serve as a surrogate for your sequence.
   

   Figure 1. Relationship among sequences, clusters, and final alignments in ProPhylER. A multiple alignment is built only from those clusters that have enough sequences to be informative for ProPhylER analyses (colored clusters). Alignment of cluster members reveals poison sequences (bad gene predictions, fragments, etc., signified by Seq1, Seq6, and Seq7) that are eliminated from the final alignment. • If you search with a sequence that is not part of a sufficiently large cluster you will get an error message to that effect. • If you search with a sequence that is part of a cluster but was eliminated from the final alignment you will still be presented with a search result.

Last updated 5/23/08