PredPPCrys: Accurate Prediction of Sequence Cloning, Protein Production, Purification and Crystallization Propensity from Protein Sequences Using Multi-Step Heterogeneous Feature Fusion and Selection

Abstract

X-ray crystallography is the primary approach to solve the three-dimensional structure of a protein. However, a major bottleneck of this method is the failure of multi-step experimental procedures to yield diffraction-quality crystals, including sequence cloning, protein material production, purification, crystallization and ultimately, structural determination. Accordingly, prediction of the propensity of a protein to successfully undergo these experimental procedures based on the protein sequence may help narrow down laborious experimental efforts and facilitate target selection. A number of bioinformatics methods based on protein sequence information have been developed for this purpose. However, our knowledge on the important determinants of propensity for a protein sequence to produce high diffraction-quality crystals remains largely incomplete. In practice, most of the existing methods display poorer performance when evaluated on larger and updated datasets. To address this problem, we constructed an up-to-date dataset as the benchmark, and subsequently developed a new approach termed ‘PredPPCrys’ using the support vector machine (SVM). Using a comprehensive set of multifaceted sequence-derived features in combination with a novel multi-step feature selection strategy, we identified and characterized the relative importance and contribution of each feature type to the prediction performance of five individual experimental steps required for successful crystallization. The resulting optimal candidate features were used as inputs to build the first-level SVM predictor (PredPPCrys I). Next, prediction outputs of PredPPCrys I were used as the input to build second-level SVM classifiers (PredPPCrys II), which led to significantly enhanced prediction performance. Benchmarking experiments indicated that our PredPPCrys method outperforms most existing procedures on both up-to-date and previous datasets. In addition, the predicted crystallization targets of currently non-crystallizable proteins were provided as compendium data, which are anticipated to facilitate target selection and design for the worldwide structural genomics consortium. PredPPCrys is freely available at http://www.structbioinfor.org/PredPPCrys.

Computational crystallization

“…This method was selected because it was found to perform better compared to support vector machine (SVM) and artificial neural network methods. Both XtalPred and XtalPred-RF have accuracies around 70%, but the latter has a higher MCC of 0.47.PPCpred [37] and PredPPCrys [51] use a SVM classifier of exposed and buried amino acid compositions, chain disorder, the proximity of certain groups of amino acids and physicochemical properties of individual and collocated amino acids. This large set of features (~800 for PPCpred and ~3000 for PredPPCrys) is then reduced to obtain a best-performing set.…”

“…A reduced set of significant features was then identified. As stated above, this method uses an earlier training dataset, and hence its prediction accuracy drops when tested with newer datasets [51]. The reported accuracy was 87%, with MCC = 0.74.…”

Protein Crystallizability

“…The authors of PredPPCrys [52] combined a novel dataset, multi-step feature selection, and SVM classification in an attempt to improve the quality of crystallization success predictions. Similar to PPCpred, their model is able to predict the success rate of individual experimental steps in the structural genomic pipeline.…”

TSTMP: target selection for structural genomics of human transmembrane proteins

“…Prediction accuracy of crystallizability prediction methods were increased by machine learning approaches such as PPCPred (14) that is based on TargetDB as well, however this application incorporated PepcDB (15) and the source data were filtered more rigorously. PredPPCrys (16) and Crysalis (17) were developed to overcome the problem of the overfitting of supervised machine learning techniques by feature selection and they were shown to be the most accurate among other methods. G-protein-coupled receptors (GPCR), for which the used structures resulted from mainly engineered proteins and short fragments that have limited usefulness for further modeling, were ordered by utilizing special propensity score (18) but it was emphasized that GPCRs are highly challenging to crystallize.…”