Abstract: 20 proteins were selected from Thermoanaerobacter tengcongensis proteome as the pilot group of targets for a structural genomics project which aims to fish for proteins with new folds and important biological functions. BLASTP, PSI-BLAST and TBLASTN searches were performed against pdb, nr and est-human databases, respectively, in order to find proteins that are least homologous to PDB proteins and most homologous to human proteins. Transmembrane protein predicting programs were used to classify membrane and soluble proteins. The selection was then made with further considerations on functions and parameters such as molecular weight and cysteine ratio. A validation made by the ProtoNet protein sequence hierarchical classification database showed that 95% of our target proteins locate in clusters that have no structures solved. Secondary structures and shapes of the target proteins predicted by PredictProtein program provided useful information for the experimental efforts. 8 selected proteins were cloned, expressed and purified, of which 2 were crystallized. Our experimental results and some recently published results show that some of our selected proteins bear new fold.

Key words: Thermoanaerobacter tengcongensis, Structural genomics, Target selection, Protein structure, BLAST, PSI-BLAST, ProtoNet