PSSM

Fig. 5. Graphical representation of an embedded PSSM representing the helix-...

The specific information present in a PSSM, even when it is made from only two or three homologous sequences, particularly favors the similarity between the sequences aligned to compute it and can improve searching sensitivity dramatically. This is because constraints on each position in a multiple alignment become better defined as more exam-... 

This overview presents some cases in which sequence profile-based methods have been able to predict nontrivial structural and evolutionary relationships between proteins and then discusses the current state of structural genomics as assesed using these methods. This discussion is not a comprehensive review of profile-based methods for sequence analysis and their application in structural genomics rather observations made with PSI-BLAST-constructed PSSMs are emphasized, and results produced by other methods are cited only as needed for discussion. 

Another important aspect of structure prediction using PSSMs is the careful determination of the boundaries of distinct domains whose sequences are used as starting points to construct PSSMs. In iterative database searches, this is critical to avoid inclusion of more than one domain, which results in explosion of an iterative search instead of convergence. Using well-defined domain sequences is particularly important in the analysis of small binding domains that frequently are overlooked if combined with larger, usually enzymatic domains. 

This section describes an example of the application of PSSMs to structure prediction that involves two previously undetected families of OB (oligomer-binding) fold domains. The OB folds were first identified... 

Detection of Unexpected Structural Relationships Among Proteins Identified Using PSI-BLAST-Constructed PSSMs... 

The first reference is for the identification of the respective domain by using PSI-BLAST-constructed PSSMs, and the other references are for similar findings made by alternative computational methods. 

The FID library was applied to the task of predicting the protein folds encoded in complete genomes using the recently developed program IMPALA, which is a modification of PSI-BLAST that effectively reverses the search protocol (Schaffer et al., 1999). PSI-BLAST compares a PSSM to a database of sequences by contrast, a single search by IMPALA is a comparison of a sequence to a library of PSSMs (Fig. 3B). Statistical tests with IMPALA have shown that the theory used for the evaluation of BLAST results is applicable with minimal modifications. 

The fold prediction rate achieved in this analysis is an even greater improvement over the results achieved with pairwise alignment methods (Fischer and Eisenberg, 1997 Gerstein and Levitt, 1997) than reported previously (Wolf et al., 1999). By contrast, several more recent studies that applied PSSM-based methods to protein sets from individual genomes have reported similar prediction rates (Huynen et al., 1998 Paw-... 

Evaluation of protein sequence analysis methods based on the use of PSSMs in benchmarking experiments and in a number of test cases shows that these methods are capable of systematically detecting relationships between proteins that previously have been deemed tractable only at the structure-comparison level. Clearly, however, there is still a lot of room for improvement, as many automated procedures missed subtle connections that subsequendy have been revealed on a case-by-case basis, in part thanks to a careful choice of starting points for the PSSM construction. An exhaustive exploration of the sequence space by recursive iterative searching is likely to yield additional, on many occasions unexpected, links between proteins and, in particular, is expected to increase the rate of structure prediction. 

Delorenzi, M., and Speed, T. (2002). An HMM model for coiled-coil domains and a comparison with PSSM-based predictions. Bioinformatics 18, 617-625. 

LA Kelley, RM MacCallum, MJE Sternberg. Enhanced genome annotation using structural profiles in the program 3D-PSSM. J Mol Biol 299 501-522, 2000. 

The 3D-PSSM (Kelley et ah, 2000) server at http //www.bmm.icnet.uk/ 3dpssm/ offers online protein fold recognition. On the submission form, enter your e-mail address and a one-line description of the query protein, then paste the query sequence into the sequence box and click the Submit button. The query sequence is used to search the Fold library for homologues. You will be informed of the URL where the result is located for 4 days. The output includes a summary table (hits with statistics models that can be viewed with RasMol classifications and links) and fold recognition by 3D-PSSM with a printout as exemplified in Figure 12.13. The alignment displays consensus sequence, secondary structures (C for coil, E for extended, and H for helix), and core score (0 for exterior to 9 for interior core). 

---