Ssu72 Method

Function

The results for the functional component of the analysis were collected from a number of sources. Firstly, a BLAST search was performed using the NCBI database and the 'blastp (protein-protein BLAST)' algorithm (1). Secondly, a literature review was conducted, using the PubMed database (2), on the hits returned from that search. Thirdly, the Drosophila Ssu72 protein, PDB accession number 3FDF_A, was submitted to the ProFunc webserver (3). The automated results returned were then analysed and further searches of PubMed were conducted as necessary - these were dictated by the structural and evolutionary results obtained in their respective sections of this paper.

Evolution

Multiple sequence alignment

NCBI’s psiBLASTp [1] search was run on all of the databases (‘nonredundant’) to yield around 200 significant matches. The search was repeated on the SwissProt database to give 24 matching high quality sequences. The Protein Data Bank yielded no significant matches.

The sequences from the ‘nonredundant’ search were aligned using ClustalX. The sequences were well conserved, so a high gap creation penalty (20) was used. Around 30 low matching sequences with large insertions and a few very short sequences were removed, and long sequences were cut to size. (Length was judged relative to the Drosophila sequence of interest.) Closely matched variants from the same organism were removed to reduce redundancy.

The sequences from the SwissProt database were also aligned using ClustalX, with a gap penalty of 20. 8 long sequences were cut to size, and redundant human sequences were removed. The Cryptococcus Neoformans sequence (Q5KIT2), derived from gene prediction, was too short to show the residues of the active site. It was replaced with a closely matching longer sequence (XM_770627.1) from RefSeq for the alignment. It was not necessary to replace the sequence for the tree, because the missing residues are so highly conserved.

Tree

The SwissProt sequences were used for the tree. Phylip-3.63 programs were used to generate the tree and bootstrap values. The tree’s structure was compared with the taxonomy provided with NCBI’s psiBLASTp results.

Comparative alignments

BLASTp was used to produce a dot matrix alignment of the target protein and its human homologue.

ClustalX was used to produce an alignment of the target protein and eight phosphatases found by Secondary Structure Matching in Profunc.

Abstract | Introduction | Results | Discussion | Method | References