SPU proteins

Table of Contents

Electronic annotation of GLEAN proteins
Manual characterization of GLEAN proteins
Alternative,complimentary classification proteins

Electronic annotation of GLEAN proteins

More than 10,000 SPU_ (GLEAN) gene models were annotated by a consortium of scientists. The remaining (approximately 18,000) SPU_ proteins were annotated as follows. First, larger proteins ( >100 amino acids) were selected, and each of them was annotated by protein sequence match. Annotated proteins from other organisms in the non-redundant protein set (nr) or reference sequences (refseq_protein) in NCBI database was used for the comparison. Initially, BLAST report (graphic output) was obtained from querying NCBI's protein data sets with each SPU_ (GLEAN) protein. Then, following criteria were used to annotate the SPU_ protein:
1. Inspect the extent of homology by surveying the span of query-to-subject sequence alignment. Typically at least 50% of query sequence alignment is required to accept the match. Once accepted, the subject gene's name and symbol are adapted to the SPU_ gene model. Usually the name is tagged by "-like" suffix. Alignments spanning at least 80% or 90% of the query sequence are regarded as true homology, and the gene name is then assigned without "-like" suffix. Gene names are assigned to SPU_ gene models in accordance to mammalian (mouse) gene names and symbols, after checking the E-value (Even if the SPU_ protein is more homologous to the proteins of insects or fungi/bacteria, we used mammalian names.)
2. An E-value greater than 50 is usually taken, but rarely, 30-50 range is also accepted if for example the query and subject sequences share critical protein motifs, or the homology occurs across a wide range of species. Usually an E-value greater than 100 or even 150 is required to assign gene names without "-like" tag.
3. Consider the uniqueness (abundance) of the SPU_ protein sequences. The more unique the sequences in the SPU_ protein (or the lower the abundance of the motifs in the SPU_ protein), the higher the clarity and confidence of the match to be valid. SPU_ proteins carrying highly redundant sequences that avidly match to dozens or hundreds of subject proteins are difficult to define.
4. Try to match the SPU_ protein to other proteins within the same S. purpurarus proteome, and consider cross-species multiple aligments also.
5. Gene models without significant homologies to proteins in other species could either be spurious proteins, or Sp-specific proteins. Usually they have paralogs with high quality matches, which then are assigned to be "hypothetical protein-xxx" where xxx is a serial number, and commented to be "Sp-specific".
6. Gene models with low quality orthologous matches are also assigned to be hypothetical protein-xxx.
7. Whenever it is possible, use gene names other than "hypothetical protein-xxx", including gene names after motif names (such as Sp-Ccdc123 or Sp-Wdr48, etc, for coiled-coil domain containing and WD repeat) or Sp-Kiaa1234 if the putative subject proteins are hypothetical, arbitrarily named proteins.
8. Redundancy of query proteins in S. purpuratus is indicated by number 1, 2, or 3 in the 9th columns of the currently used gene_info_table format. Pol polyprotein-like (presumably from retroposons), ankyrin-repeat sequences, notch-like or EGF-like sequences occur most frequently, and serial numbers are used to name these genes.

Back to the top of the page

Manual characterization of GLEAN proteins

Back to the top of the page

Alternative,complimentary classification proteins

We are currently working on a new scheme to classify Sp proteome into alternative protein groups. The scheme is intended to be radically different from existing protein classification approaches, and aims to provide database query terms that are complementary to conventional protein family/domain names. 3,362 Pfam domain names from the GLEAN proteins were re-grouped according to a variety of attributes including structure, biological processes, intracellular association, enzymatic activity, etc. Currently 3,362 pfam groups have been compressed to 135 new groups.
http://piano.caltech.edu/4/135-alternative_categories
http://piano.caltech.edu/4/alternative_category-pfam.correlationtable

Back to the top of the page

Search form

SPU proteins

Electronic annotation of GLEAN proteins

Manual characterization of GLEAN proteins

Alternative,complimentary classification proteins