Modeling protein families using probabilistic suffix trees

Abstract

proteins which await analysis.We present a method for modeling protein families by means of probabilistic suffix trees (PSTs). The method is based on identifying significant patterns in a set of related protein sequences. The input sequences do not need to be aligned, nor is delineation of domain boundaries required. The method is automatic, and can be applied, without assuming any preliminary biological information, with surprising success. Incorporating basic biological considerations such as amino acid background probabilities, and amino acids substitution probabilities can improve the performance in some cases.The PST can serve as a predictive tool for protein sequence classification, and for detecting conserved patterns (possibly functionally or structurally important) within protein sequences. The method was tested on one of the state of the art databases of protein families, namely, the Pfam database of HMMs, with satisfactory performance.

Protein Sequence Classification Through Relevant Sequence Mining and Bayes Classifiers

“…The classification result is determined by equation 4. The evaluation in [11] and [8] was different from ours. They used 4/5 of the family sequences to build a model for the respective family and evaluated the model with the remaining 1/5 of the sequences.…”

“…Table 2 shows the average classification results for table 1, when all the families are considered (row 2), when the family PPR is left out (row 3) and when PPR and TPR are both left out of the classification (row 4). In table 3 we compare the three probabilistic models with the results from the PSTs and SMTs published in [11,8]. In the last row of the table we present the average classification results.…”

“…-The Probabilistic Model Classifiers, that work by simulating the sequence family under consideration. Typical probabilistic classifiers are the simple and k-order Markov Chain [19], Hidden Markon Models [14] and Probabilistic Suffix Trees [11].…”

“…Recently Probabilistic Suffix Trees (PSTs) [11] and Sparse Markov Transducers (SMTs) [8] have been applied in the protein classification problem, and have shown superior performance. A PST is essentially a variable length Markov Model, where the probability of a symbol in a sequence depends on the previous symbols.…”

“…A wild-card is a symbol that denotes a gap of size one and matches any symbol on the alphabet. In [11] an experimental evaluation has shown that PSTs perform much better than a typical BLAST search and as good as HMM. This is very interesting since the latter approach makes use of multiple alignments and the families are usually defined based on an HMM [9].…”