Structure prediction, from sequence

Currently, methods of structure prediction from sequence perform poorly. The results of the biannual CASP experiment (Critical Assessment of techniques for protein Structure Prediction see http //predictioncenter.llnl. gov/) are equivocal to say the least. A recent report on the improvements of aligning target sequences to a structural template (67) indicates that over the last four CASP competitions there was no significant improvement in quality in this key step in the prediction process. Alignment remains the major source of error in all models based on less than 30% se-... 

STRUCTURE PREDICTION FROM SEQUENCE BY STATISTICAL METHODS... 

There is a considerable impetus to predict accurately protein structures from sequence information because of the protein sequence/structure deficit as a consequence of the genome and full-length cDNA sequencing projects. The molecular mechanical (MM) approach to modeling of protein structures has been discussed in section 9.2, and the protein secondary structure prediction from sequence by statistical methods has been treated in section 9.5. The prediction of protein structure using bioinformatic resources will be described in this subsection. The approaches to protein structure predictions from amino acid sequences (Tsigelny, 2002 Webster, 2000) include ... 

Only a few residues show more than 75% sequence identity, including four glycine residues, a proline residue at the beginning of the Pro loop, and a phenylalanine residue in a position corresponding to the conserved residue Tyr 165 of the bovine heart Rieske protein. However, structure prediction and sequence comparison with Rieske proteins from bci complexes suggests that the fold will be very similar in all Rieske-type ferredoxins, as in the other Rieske or Rieske-type proteins (see Section III,B,1). 

Challis GL, Ravel J (2000) Coelichelin, a New Peptide Siderophore Encoded by the Streptomyces coelicolor genome Structure Prediction from the Sequence of its Non-Ribosomal Peptide Synthetase. FEMS Microbiol Lett 187 111... 

Earlier structure predictions with coronins resulted in models of five-bladed P propellers, because only five WD repeats were recognised. However, many p propeller structures contain WD repeats with litde or no obvious sequence homology to the canonical sequence. Additionally, while typical WD repeats possess four to eight copies and up to 16 copies have been reported, almost all structures ofWD repeat proteins in the PDB contain seven blades per propeller. Keeping this in mind, more detailed analyses of coronins have been carried out and additional WD-like repeats were identified based on secondary structure prediction from amino acid sequences, indicating the presence of seven copies for coronins I(IA) - and 3 (IC) and 14 for coronin 7. The fold of a seven-bladed propeller was then confirmed by the crystal structure of a coronin 1 (lA) fragment. ... 

In the absence of a high-resolution three-dimensional structure of a BFR, a tentative subunit conformation (Fig. 10) has been derived by secondary structure prediction from amino acid sequence data. A combined prediction based on eight different programs gave a helical dis-... 

In contrast to rational design methods, where changes in protein properties may be predicted from sequence changes, random mutation methods have been introduced to empirically improve protein properties in the absence of a priori knowledge of structure-function relationships. Error-prone polymerase chain reaction (PCR), for example, can be used to introduce approximately random mutations in amplified DNA sequences. The protein products of these mutations may then be screened to select successful or improved variants.2... 

Frishman, D. and Argos, R, Incorporation of non-local interactions in protein secondary structure prediction from the amino acid sequence, Prot. Eng., 9, 133, 1996. 

Edwards, Y. J. K. and Perkins, S. J., Assessment of protein fold predictions from sequence information the predicted a/ 3 doubly wound fold of the van WUlebrand factor type A domain is similar to its crystal structure, J. Mol Biol, 260, 277, 1996. 

ML techniques are well situated for the analysis of molecular sequence data. These methods have been applied successfully to a variety problem, ranging from gene identification to protein structure prediction and sequence classihcation (24,25). These techniques have become an important topic for... 

Figure 4 PHDsec profile-based neural network system for secondary structure prediction. From the multiple alignment (here guide sequence SH3 plus four other proteins al -a4, note lower case letters indicate deletions in the aligned sequence) a profile of amino acid occurrences is compiled. To the resulting 20 values at one particular position p in the protein (one column) three values are added the number of deletions and insertions, and the conservation weight (CW). 13 adjacent columns are used as input. The whole network system for. secondary structure prediction consists of three layers two network layers and one layer averaging over independently trained networks...

---