A graph-based clustering method applied to protein sequences

The number of amino acid sequences is increasing very rapidly in the protein databases like Swiss-Prot, Uniprot, PIR and others, but the structure of only some amino acid sequences are found in the Protein Data Bank. Thus, an important problem in genomics is automatically clustering homologous protein sequences when only sequence information is available. Here, we use graph theoretic techniques for clustering amino acid sequences. A similarity graph is defined and clusters in that graph correspond to connected subgraphs. Cluster analysis seeks grouping of amino acid sequences into subsets based on distance or similarity score between pairs of sequences. Our goal is to find disjoint subsets, called clusters, such that two criteria are satisfied: homogeneity: sequences in the same cluster are highly similar to each other; and separation: sequences in different clusters have low similarity to each other. We tested our method on several subsets of SCOP (Structural Classification of proteins) database, a gold standard for protein structure classification. The results show that for a given set of proteins the number of clusters we obtained is close to the superfamilies in that set; there are fewer singeltons; and the method correctly groups most remote homologs.     

Role of thymine in protein coding frames of mRNA sequences

Distribution of thymine in protein coding mRNA sequences has been studied here. Our study suggest that thymine in protein coding sequences are not randomly distributed but with probability. Frame1 prefers to have definite amount of thymine. It is observed that the thymine content of frame 4 is also involved in protein coding. Frame 3 prefers to have least amount of thymine. However, frame 2 and frame 6 shows a variable degree of thymine content. The mRNA sequences of heterosexual animals, particularly, the human show a different distribution profile (less thymine in frame 1) compared to that of yeast and plants.     

Artificial protein sequences enable recognition of vicinal and distant protein functional relationships

High divergence in protein sequences makes the detection of distant protein relationships through homology-based approaches challenging. Grouping protein sequences into families, through similarities in either sequence or 3-D structure, facilitates in the improved recognition of protein relationships. In addition, strategically designed protein-like sequences have been shown to bridge distant structural domain families by serving as artificial linkers. In this study, we have augmented a search database of known protein domain families with such designed sequences, with the intention of providing functional clues to domain families of unknown structure. When assessed using representative query sequences from each family, we obtain a success rate of 94% in protein domain families of known structure. Further, we demonstrate that the augmented search space enabled fold recognition for 582 families with no structural information available a priori. Additionally, we were able to provide reliable functional relationships for 610 orphan families. We discuss the application of our method in predicting functional roles through select examples for DUF4922, DUF5131, and DUF5085. Our approach also detects new associations between families that were previously not known to be related, as demonstrated through new sub-groups of the RNA polymerase domain among three distinct RNA viruses. Taken together, designed sequences-augmented search databases direct the detection of meaningful relationships between distant protein families. In turn, they enable fold recognition and offer reliable pointers to potential functional sites that may be probed further through direct mutagenesis studies.     

Swaps in protein sequences

An important question in protein evolution is to what extent proteins may have undergone swaps (switches of domain or fragment order) during evolution. Such events might have occurred in several forms: Swaps of short fragments, swaps of structural and functional motifs, or recombination of domains in multidomain proteins. This question is important for the theoretical understanding of the evolution of proteins, and has practical implications for using swaps as a design tool in protein engineering. In order to analyze the question systematically, we conducted a large scale survey of possible swaps and permutations among all pairs of protein from the Swissport database. A swap is defined as a specific kind of sequence mutation between two proteins in which two fragments that appear in both sequences have different relative order in the two sequences. For example, aXbYc and dYeXf are defined as a swap, where X and Y represent sequence fragments that switched their order. Identifying such swaps is difficult using standard sequence comparison packages. One of the main problems in the analysis stems from the fact that many sequences contain repeats, which may be identified as false-positive swaps. We have used two different approaches to detect pairs of proteins with swaps. The first approach is based on the predefined list of domains in Pfam. We identified all the proteins that share at least two domains and analyzed their relative order, looking for pairs in which the order of these domains was switched. We designed an algorithm to distinguish between real swaps and duplications. In the second approach, we used Blast to detect pairs of proteins that share several fragments. Then, we used an automatic procedure to select pairs that are likely to contain swaps. Those pairs were analyzed visually, using a graphical tool, to eliminate duplications. Combining these approaches, about 140 different cases of swaps in the Swissprot database were found (after eliminating multiple pairs within the same family). Some of the cases have been described in the literature, but many are novel examples. Although each new example identified may be interesting to analyze, our main conclusion is that cases of swaps are rare in protein evolution. This observation is at odds with the common view that proteins are very modular to the point that modules (e.g., domains) can be shuffled between proteins with minimal constraints. Our study suggests that sequential constraints, i.e., the relative order between domains, are highly conserved.     

Alignment of protein sequences by their profiles

The accuracy of an alignment between two protein sequences can be improved by including other detectably related sequences in the comparison. We optimize and benchmark such an approach that relies on aligning two multiple sequence alignments, each one including one of the two protein sequences. Thirteen different protocols for creating and comparing profiles corresponding to the multiple sequence alignments are implemented in the SALIGN command of MODELLER. A test set of 200 pairwise, structure-based alignments with sequence identities below 40% is used to benchmark the 13 protocols as well as a number of previously described sequence alignment methods, including heuristic pairwise sequence alignment by BLAST, pairwise sequence alignment by global dynamic programming with an affine gap penalty function by the ALIGN command of MODELLER, sequence-profile alignment by PSI-BLAST, Hidden Markov Model methods implemented in SAM and LOBSTER, pairwise sequence alignment relying on predicted local structure by SEA, and multiple sequence alignment by CLUSTALW and COMPASS. The alignment accuracies of the best new protocols were significantly better than those of the other tested methods. For example, the fraction of the correctly aligned residues relative to the structure-based alignment by the best protocol is 56%, which can be compared with the accuracies of 26%, 42%, 43%, 48%, 50%, 49%, 43%, and 43% for the other methods, respectively. The new method is currently applied to large-scale comparative protein structure modeling of all known sequences.     

Self-organized neural maps of human protein sequences.

We have recently described a method based on artificial neural networks to cluster protein sequences into families. The network was trained with Kohonen''s unsupervised learning algorithm using, as inputs, the matrix patterns derived from the dipeptide composition of the proteins. We present here a large-scale application of that method to classify the 1,758 human protein sequences stored in the SwissProt database (release 19.0), whose lengths are greater than 50 amino acids. In the final 2-dimensional topologically ordered map of 15 x 15 neurons, proteins belonging to known families were associated with the same neuron or with neighboring ones. Also, as an attempt to reduce the time-consuming learning procedure, we compared 2 learning protocols: one of 500 epochs (100 SUN CPU-hours [CPU-h]), and another one of 30 epochs (6.7 CPU-h). A further reduction of learning-computing time, by a factor of about 3.3, with similar protein clustering results, was achieved using a matrix of 11 x 11 components to represent the sequences. Although network training is time consuming, the classification of a new protein in the final ordered map is very fast (14.6 CPU-seconds). We also show a comparison between the artificial neural network approach and conventional methods of biosequence analysis.     

Selection against toxic aggregation-prone protein sequences in bacteria

Despite genetic variation has the potential to arise new protein functions, spontaneous mutations usually destabilize the native fold. Misfolded proteins tend to form cytotoxic intracellular aggregates, decreasing cell fitness and leading to degenerative disorders in humans. Therefore, it is thought that selection against protein misfolding and aggregation constrains the evolution of protein sequences. However, obtaining experimental data to validate this hypothesis has been traditionally difficult. Here we exploit bacteria as a model organism to address this question. Using variants of the Alzheimer's related Aβ42 peptide designed to exhibit different in vivo aggregation propensities we show here that, in cell competition experiments, the most aggregation-prone variants are always purged out from the growing population. Flow cytometry analysis of cellular metabolism and viability demonstrates that this purifying effect responds to a clear correlation between physiological burden and intrinsic aggregation propensity. Interestingly, the fitness cost of aggregation appears to be associated with aggregation rates rather than with overall protein solubility. Accordingly, we show that, by reducing in vivo aggregation rates, the model osmolyte proline is able to buffer the metabolic impact of protein aggregation. Overall, our data provide experimental support for the role of toxic protein aggregation on the cell fitness landscape and the evolution of natural protein sequences.     

Recognition of distantly related protein sequences using conserved motifs and neural networks.

A sensitive technique for protein sequence motif recognition based on neural networks has been developed. It involves three major steps. (1) At each appropriate alignment position of a set of N matched sequences, a set of N aligned oligopeptides is specified with preselected window length. N neural nets are subsequently and successively trained on N-1 amino acid spans after eliminating each ith oligopeptide. A test for recognition of each of the ith spans is performed. The average neural net recognition over N such trials is used as a measure of conservation for the particular windowed region of the multiple alignment. This process is repeated for all possible spans of given length in the multiple alignment. (2) The M most conserved regions are regarded as motifs and the oligopeptides within each are used to train intensively M individual neural networks. (3) The M networks are then applied in a search for related primary structures in a databank of known protein sequences. The oligopeptide spans in the database sequence with strongest neural net output for each of the M networks are saved and then scored according to the output signals and the proper combination that follows the expected N- to C-terminal sequence order. The motifs from the database with highest similarity scores can then be used to retrain the M neural nets, which can be subsequently utilized for further searches in the databank, thus providing even greater sensitivity to recognize distant familial proteins. This technique was successfully applied to the integrase, DNA-polymerase and immunoglobulin families.     

Deciphering membrane protein structures from protein sequences

ABSTRACT: Co-evolving positions within protein sequences have been used as spatial constraints to develop a computational approach for modeling membrane protein structures.     

Clustering of multi‐domain protein sequences

The overall function of a multi‐domain protein is determined by the functional and structural interplay of its constituent domains. Traditional sequence alignment‐based methods commonly utilize domain‐level information and provide classification only at the level of domains. Such methods are not capable of taking into account the contributions of other domains in the proteins, and domain‐linker regions and classify multi‐domain proteins. An alignment‐free protein sequence comparison tool, CLAP (CLAssification of Proteins) was previously developed in our laboratory to especially handle multi‐domain protein sequences without a requirement of defining domain boundaries and sequential order of domains. Through this method we aim to achieve a biologically meaningful classification scheme for multi‐domain protein sequences. In this article, CLAP‐based classification has been explored on 5 datasets of multi‐domain proteins and we present detailed analysis for proteins containing (1) Tyrosine phosphatase and (2) SH3 domain. At the domain‐level CLAP‐based classification scheme resulted in a clustering similar to that obtained from an alignment‐based method. CLAP‐based clusters obtained for full‐length datasets were shown to comprise of proteins with similar functions and domain architectures. Our study demonstrates that multi‐domain proteins could be classified effectively by considering full‐length sequences without a requirement of identification of domains in the sequence.     

Role of low-complexity sequences in the formation of novel protein coding sequences

Low-complexity sequences are extremely abundant in eukaryotic proteins for reasons that remain unclear. One hypothesis is that they contribute to the formation of novel coding sequences, facilitating the generation of novel protein functions. Here, we test this hypothesis by examining the content of low-complexity sequences in proteins of different age. We show that recently emerged proteins contain more low-complexity sequences than older proteins and that these sequences often form functional domains. These data are consistent with the idea that low-complexity sequences may play a key role in the emergence of novel genes.     

Information content of protein sequences

The complexity of large sets of non-redundant protein sequences is measured. This is done by estimating the Shannon entropy as well as applying compression algorithms to estimate the algorithmic complexity. The estimators are also applied to randomly generated surrogates of the protein data. Our results show that proteins are fairly close to random sequences. The entropy reduction due to correlations is only about 1%. However, precise estimations of the entropy of the source are not possible due to finite sample effects. Compression algorithms also indicate that the redundancy is in the order of 1%. These results confirm the idea that protein sequences can be regarded as slightly edited random strings. We discuss secondary structure and low-complexity regions as causes of the redundancy observed. The findings are related to numerical and biochemical experiments with random polypeptides.     

Oligopeptide repeats in the yeast protein Sup35p stabilize intermolecular prion interactions

The nuclear-encoded Sup35p protein is responsible for the prion-like [PSI(+)] determinant of yeast, with Sup35p existing largely as a high molecular weight aggregate in [PSI(+)] strains. Here we show that the five oligopeptide repeats present at the N-terminus of Sup35p are responsible for stabilizing aggregation of Sup35p in vivo. Sequential deletion of the oligopeptide repeats prevented the maintenance of [PSI(+)] by the truncated Sup35p, although deletants containing only two repeats could be incorporated into pre-existing aggregates of wild-type Sup35p. The mammalian prion protein PrP also contains similar oligopeptide repeats and we show here that a human PrP repeat (PHGGGWGQ) is able functionally to replace a Sup35p oligopeptide repeat to allow stable [PSI(+)] propagation in vivo. Our data suggest a model in which the oligopeptide repeats in Sup35p stabilize intermolecular interactions between Sup35p proteins that initiate establishment of the aggregated state. Modulating repeat number therefore alters the rate of yeast prion conversion in vivo. Furthermore, there appears to be evolutionary conservation of function of the N-terminally located oligopeptide repeats in prion propagation.     

Detection of native-like models for amino acid sequences of unknown three-dimensional structure in a data base of known protein conformations.

We present an approach which can be used to identify native-like folds in a data base of protein conformations in the absence of any sequence homology to proteins in the data base. The method is based on a knowledge-based force field derived from a set of known protein conformations. A given sequence is mounted on all conformations in the data base and the associated energies are calculated. Using several conformations and sequences from the globin family we show that the native conformation is identified correctly. In fact the resolution of the force field is high enough to discriminate between a native fold and several closely related conformations. We then apply the procedure to several globins of known sequence but unknown three dimensional structure. The homology of these sequences to globins of known structures in the data base ranges from 49 to 17%. With one exception we find that for all globin sequences one of the known globin folds is identified as the most favorable conformation. These results are obtained using a force field derived from a data base devoid of globins of known structure. We briefly discuss useful applications in protein structural research and future development of our approach.     

Frequent in-frame length variations are found in the diverged simple repeat sequences of the protein-coding regions of two putative protein kinase genes of Brassica napus

Two putative protein kinase cDNA clones were isolated from Brassica napus by screening with a putative protein kinase cDNA clone of Arabidopsis thaliana. The deduced amino acid sequences show a distinct modular composition, consisting of a possible protein kinase catalytic region at the amino terminus and a highly acidic region encoded from diverged simple repeat sequences at the carboxy terminus. Comparison of the nucleotide sequences encoding this acidic region revealed a high rate of in-frame length variation, while preserving the acidic characteristics. Similar variation is also found in the non-coding regions of these clones.     

Topological maps of protein sequences

A new method based on neural networks to cluster proteins into families is described. The network is trained with the Kohonen unsupervised learning algorithm, using matrix pattern representations of the protein sequences as inputs. The components (x, y) of these 20×20 matrix patterns are the normalized frequencies of all pairs xy of amino acids in each sequence. We investigate the influence of different learning parameters in the final topological maps obtained with a learning set of ten proteins belonging to three established families. In all cases, except in those where the synaptic vectors remains nearly unchanged during learning, the ten proteins are correctly classified into the expected families. The classification by the trained network of mutated or incomplete sequences of the learned proteins is also analysed. The neural network gives a correct classification for a sequence mutated in 21.5%±7% of its amino acids and for fragments representing 7.5%±3% of the original sequence. Similar results were obtained with a learning set of 32 proteins belonging to 15 families. These results show that a neural network can be trained following the Kohonen algorithm to obtain topological maps of protein sequences, where related proteins are finally associated to the same winner neuron or to neighboring ones, and that the trained network can be applied to rapidly classify new sequences. This approach opens new possibilities to find rapid and efficient algorithms to organize and search for homologies in the whole protein database.     

Searching for target sequences by p53 protein is influenced by DNA length

One of the most important functions of the tumor suppressor p53 protein is its sequence-specific binding to DNA. Using a competition assay on agarose gels we found that the p53 consensus sequences in longer DNA fragments are better targets than the same sequences in shorter DNAs. Semi-quantitative evaluation of the competition experiments showed a correlation between the relative p53-DNA binding and the DNA lengths. Our results are consistent with a model of the p53-DNA interactions involving one-dimensional migration of the p53 protein along the DNA for distances of about 1000 bp while searching for its target sites. Positioning of the p53 target in the DNA fragment did not substantially affect the apparent p53-DNA binding, suggesting that p53 can slide along the DNA in a bi-directional manner. In contrast to full-length p53, the isolated core domain did not show any significant correlation between sequence-specific DNA binding and fragment length.     

A systematic search for protein signature sequences.

Signature sequences are contiguous patterns of amino acids 10-50 residues long that are associated with a particular structure or function in proteins. These may be of three types (by our nomenclature): superfamily signatures, remnant homologies, and motifs. We have performed a systematic search through a database of protein sequences to automatically and preferentially find remnant homologies and motifs. This was accomplished in three steps: 1. We generated a nonredundant sequence database. 2. We used BLAST3 (Altschul and Lipman, Proc. Natl. Acad. Sci. U.S.A. 87:5509-5513, 1990) to generate local pairwise and triplet sequence alignments for every protein in the database vs. every other. 3. We selected "interesting" alignments and grouped them into clusters. We find that most of the clusters contain segments from proteins which share a common structure or function. Many of them correspond to signatures previously noted in the literature. We discuss three previously recognized motifs in detail (FAD/NAD-binding, ATP/GTP-binding, and cytochrome b5-like domains) to demonstrate how the alignments generated by our procedure are consistent with previous work and make structural and functional sense. We also discuss two signatures (for N-acetyltransferases and glycerol-phosphate binding) which to our knowledge have not been previously recognized.