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

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.     

Protein biophysical properties that correlate with crystallization success in Thermotoga maritima: maximum clustering strategy for structural genomics

Cost and time reduction are two of the driving forces in the development of new strategies for protein crystallization and subsequent structure determination. Here, we report the analysis of the Thermotoga maritima proteome, in which we compare the proteins that were successfully expressed, purified and crystallized versus the rest of the proteome. This set of almost 500 proteins represents one of the largest, internally consistent, protein expression and crystallization datasets available. The analysis shows that individual parameters, such as isoelectric point, sequence length, average hydropathy, low complexity regions (SEG), and combinations of these biophysical properties for crystallized proteins define a distinct subset of the T. maritima proteome. The distribution profiles of the various biophysical properties in the expression/crystallization set are then used to extract rules to improve target selection and improve the efficiency and output of structural genomics, as well as general structural biology efforts.     

The challenge of protein structure determination--lessons from structural genomics

The process of experimental determination of protein structure is marred with a high ratio of failures at many stages. With availability of large quantities of data from high-throughput structure determination in structural genomics centers, we can now learn to recognize protein features correlated with failures; thus, we can recognize proteins more likely to succeed and eventually learn how to modify those that are less likely to succeed. Here, we identify several protein features that correlate strongly with successful protein production and crystallization and combine them into a single score that assesses "crystallization feasibility." The formula derived here was tested with a jackknife procedure and validated on independent benchmark sets. The "crystallization feasibility" score described here is being applied to target selection in the Joint Center for Structural Genomics, and is now contributing to increasing the success rate, lowering the costs, and shortening the time for protein structure determination. Analyses of PDB depositions suggest that very similar features also play a role in non-high-throughput structure determination, suggesting that this crystallization feasibility score would also be of significant interest to structural biology, as well as to molecular and biochemistry laboratories.     

XANNpred: neural nets that predict the propensity of a protein to yield diffraction-quality crystals

Production of diffracting crystals is a critical step in determining the three-dimensional structure of a protein by X-ray crystallography. Computational techniques to rank proteins by their propensity to yield diffraction-quality crystals can improve efficiency in obtaining structural data by guiding both protein selection and construct design. XANNpred comprises a pair of artificial neural networks that each predict the propensity of a selected protein sequence to produce diffraction-quality crystals by current structural biology techniques. Blind tests show XANNpred has accuracy and Matthews correlation values ranging from 75% to 81% and 0.50 to 0.63 respectively; values of area under the receiver operator characteristic (ROC) curve range from 0.81 to 0.88. On blind test data XANNpred outperforms the other available algorithms XtalPred, PXS, OB-Score, and ParCrys. XANNpred also guides construct design by presenting graphs of predicted propensity for diffraction-quality crystals against residue sequence position. The XANNpred-SG algorithm is likely to be most useful to target selection in structural genomics consortia, while the XANNpred-PDB algorithm is more suited to the general structural biology community. XANNpred predictions that include sliding window graphs are freely available from http://www.compbio.dundee.ac.uk/xannpred     

Evaluation of features for catalytic residue prediction in novel folds

Structural genomics projects are determining the three-dimensional structure of proteins without full characterization of their function. A critical part of the annotation process involves appropriate knowledge representation and prediction of functionally important residue environments. We have developed a method to extract features from sequence, sequence alignments, three-dimensional structure, and structural environment conservation, and used support vector machines to annotate homologous and nonhomologous residue positions based on a specific training set of residue functions. In order to evaluate this pipeline for automated protein annotation, we applied it to the challenging problem of prediction of catalytic residues in enzymes. We also ranked the features based on their ability to discriminate catalytic from noncatalytic residues. When applying our method to a well-annotated set of protein structures, we found that top-ranked features were a measure of sequence conservation, a measure of structural conservation, a degree of uniqueness of a residue's structural environment, solvent accessibility, and residue hydrophobicity. We also found that features based on structural conservation were complementary to those based on sequence conservation and that they were capable of increasing predictor performance. Using a family nonredundant version of the ASTRAL 40 v1.65 data set, we estimated that the true catalytic residues were correctly predicted in 57.0% of the cases, with a precision of 18.5%. When testing on proteins containing novel folds not used in training, the best features were highly correlated with the training on families, thus validating the approach to nonhomologous catalytic residue prediction in general. We then applied the method to 2781 coordinate files from the structural genomics target pipeline and identified both highly ranked and highly clustered groups of predicted catalytic residues.     

Crystallization data mining in structural genomics: using positive and negative results to optimize protein crystallization screens

Recent efforts to collect and mine crystallization data from structural genomics (SG) consortia have led to the identification of minimal screens and novel screening strategies that can be used to streamline the crystallization process. Two groups, the Joint Center for Structural Genomics and the University of Toronto, carried out large-scale crystallization trials on different sets of bacterial targets (539, JCSG and 755, Toronto), using different sample processing and crystallization methods, and then analyzed their results to identify the smallest subset of conditions that would have crystallized the maximum number of protein targets. The JCSG Core Screen contains 67 conditions (from 480) while the Toronto Minimal Screen contains 6 (from 48). While the exact conditions included in the two screens do not overlap, the major precipitants of the conditions are similar and thus both screens can be used to determine if a protein has a natural propensity to crystallize. In addition, studies from other groups including the University of Queensland, the Mycobacterium tuberculosis SG group, the Southeast Collaboratory for SG, and the York Structural Biology Laboratory indicate that alternative crystallization strategies may be more successful at identifying initial crystallization conditions than typical sparse matrix screens. These minimal screens and alternative screening strategies are already being used to optimize the crystallization processes within large SG efforts. The differences between these results, however, demonstrate that additional studies which examine the influence of protein biophysical properties and sample preparation methods on crystal formation must also be carried out before more robust screens can be identified.     

Structural Genomics of Eukaryotic Targets at a Laboratory Scale

Structural genomics programs are distributed worldwide and funded by large institutions such as the NIH in United-States, the RIKEN in Japan or the European Commission through the SPINE network in Europe. Such initiatives, essentially managed by large consortia, led to technology and method developments at the different steps required to produce biological samples compatible with structural studies. Besides specific applications, method developments resulted mainly upon miniaturization and parallelization. The challenge that academic laboratories faces to pursue structural genomics programs is to produce, at a higher rate, protein samples. The Structural Biology and Genomics Department (IGBMC – Illkirch – France) is implicated in a structural genomics program of high eukaryotes whose goal is solving crystal structures of proteins and their complexes (including large complexes) related to human health and biotechnology. To achieve such a challenging goal, the Department has established a medium-throughput pipeline for producing protein samples suitable for structural biology studies. Here, we describe the setting up of our initiative from cloning to crystallization and we demonstrate that structural genomics may be manageable by academic laboratories by strategic investments in robotic and by adapting classical bench protocols and new developments, in particular in the field of protein expression, to parallelization.     

Protein crystallization: virtual screening and optimization

Advances in genomics have yielded entire genetic sequences for a variety of prokaryotic and eukaryotic organisms. This accumulating information has escalated the demands for three-dimensional protein structure determinations. As a result, high-throughput structural genomics has become a major international research focus. This effort has already led to several significant improvements in X-ray crystallographic and nuclear magnetic resonance methodologies. Crystallography is currently the major contributor to three-dimensional protein structure information. However, the production of soluble, purified protein and diffraction-quality crystals are clearly the major roadblocks preventing the realization of high-throughput structure determination.

This paper discusses a novel approach that may improve the efficiency and success rate for protein crystallization. An automated nanodispensing system is used to rapidly prepare crystallization conditions using minimal sample. Proteins are subjected to an incomplete factorial screen (balanced parameter screen), thereby efficiently searching the entire “crystallization space” for suitable conditions. The screen conditions and scored experimental results are subsequently analyzed using a neural network algorithm to predict new conditions likely to yield improved crystals. Results based on a small number of proteins suggest that the combination of a balanced incomplete factorial screen and neural network analysis may provide an efficient method for producing diffraction-quality protein crystals.     

Protein crystallization in the structural genomics era

There are five broad areas where noteworthy advances have occurred in the field of macromolecular crystallization in the past 10 years, though some areas have seen the major part of those advances in only the last two years. This is largely a consequence of the international structural genomics initiative and its early results. The five areas are: (1) Physical studies and characterization of the protein crystallization process; (2) Development of new practical approaches and procedures; (3) The implementation of protein engineering by genetic means to enhance both purification and crystallization; (4) The creation of new screening conditions based on information and databases emerging from structural genomics; and (5) Development and implementation of automation, robotics, and mass screening of crystallization conditions using very small amounts of protein. A brief summary is provided here of the progress in the past few years and the influence of the structural genomics project.     

Identifying foldable regions in protein sequence from the hydrophobic signal

Structural genomics initiatives aim to elucidate representative 3D structures for the majority of protein families over the next decade, but many obstacles must be overcome. The correct design of constructs is extremely important since many proteins will be too large or contain unstructured regions and will not be amenable to crystallization. It is therefore essential to identify regions in protein sequences that are likely to be suitable for structural study. Scooby-Domain is a fast and simple method to identify globular domains in protein sequences. Domains are compact units of protein structure and their correct delineation will aid structural elucidation through a divide-and-conquer approach. Scooby-Domain predictions are based on the observed lengths and hydrophobicities of domains from proteins with known tertiary structure. The prediction method employs an A*-search to identify sequence regions that form a globular structure and those that are unstructured. On a test set of 173 proteins with consensus CATH and SCOP domain definitions, Scooby-Domain has a sensitivity of 50% and an accuracy of 29%, which is better than current state-of-the-art methods. The method does not rely on homology searches and, therefore, can identify previously unknown domains.     

Thermofluor-based optimization strategy for the stabilization and crystallization of Campylobacter jejuni desulforubrerythrin

Desulforubrerythrin from Campylobacter jejuni has recently been biochemical and spectroscopically characterized. It is a member of the rubrerythrin family, and it is composed of three structural domains: the N-terminal desulforedoxin domain with a non-heme iron center, followed by a four-helix bundle domain harboring a binuclear iron center and finally a C-terminal rubredoxin domain. To date, this is the first example of a protein presenting this kind of structural domain organization, and therefore the determination of its crystal structure may unveil unexpected structural features. Several attempts were made in order to obtain protein crystals, but always without success. As part of our strategy the thermofluor method was used to increase protein stability and its propensity to crystallize. This approach has been recently used to optimize protein buffer formulation, thus yielding more stable and homogenous protein samples. Thermofluor has also been used to identify cofactors/ligands or small molecules that may help stabilize native protein states. A successful thermofluor approach was used to select a pH buffer condition that allowed the crystallization of Campylobacter jejuni desulforubrerythrin, by screening both buffer pH and salt concentration. A buffer formulation was obtained which increased the protein melting temperature by 7°C relatively to the initial purification buffer. Desulforubrerythrin was seen to be stabilized by lower pH and high salt concentration, and was dialyzed into the new selected buffer, 100mM MES pH 6.2, 500mM NaCl. This stability study was complemented with a second thermofluor assay in which different additives were screened. A crystallization screening was carried out and protein crystals were rapidly obtained in one condition. Protein crystal optimization was done using the same additive screening. Interestingly, a correlation between the stability studies and crystallization experiments using the additive screening could be established. The work presented here shows an elegant example where thermofluor was shown to be a key biophysical method that allowed the identification of an improved buffer formulation and the applicability of this technique to increase the propensity of a protein to crystallize is discussed.     

The use of soluble protein structures in modeling helical proteins in a layered membrane

Major advances have been made in the prediction of soluble protein structures, led by the knowledge-based modeling methods that extract useful structural trends from known protein structures and incorporate them into scoring functions. The same cannot be reported for the class of transmembrane proteins, primarily due to the lack of high-resolution structural data for transmembrane proteins, which render many of the knowledge-based method unreliable or invalid. We have developed a method that harnesses the vast structural knowledge available in soluble protein data for use in the modeling of transmembrane proteins. At the core of the method, a set of transmembrane protein decoy sets that allow us to filter and train features recognized from soluble proteins for transmembrane protein modeling into a set of scoring functions. We have demonstrated that structures of soluble proteins can provide significant insight into transmembrane protein structures. A complementary novel two-stage modeling/selection process that mimics the two-stage helical membrane protein folding was developed. Combined with the scoring function, the method was successfully applied to model 5 transmembrane proteins. The root mean square deviations of the predicted models ranged from 5.0 to 8.8?Å to the native structures.     

FrankSum: new feature selection method for protein function prediction

In the study of in silico functional genomics, improving the performance of protein function prediction is the ultimate goal for identifying proteins associated with defined cellular functions. The classical prediction approach is to employ pairwise sequence alignments. However this method often faces difficulties when no statistically significant homologous sequences are identified. An alternative way is to predict protein function from sequence-derived features using machine learning. In this case the choice of possible features which can be derived from the sequence is of vital importance to ensure adequate discrimination to predict function. In this paper we have successfully selected biologically significant features for protein function prediction. This was performed using a new feature selection method (FrankSum) that avoids data distribution assumptions, uses a data independent measurement (p-value) within the feature, identifies redundancy between features and uses an appropriate ranking criterion for feature selection. We have shown that classifiers generated from features selected by FrankSum outperforms classifiers generated from full feature sets, randomly selected features and features selected from the Wrapper method. We have also shown the features are concordant across all species and top ranking features are biologically informative. We conclude that feature selection is vital for successful protein function prediction and FrankSum is one of the feature selection methods that can be applied successfully to such a domain.     

Predicting protein crystallization propensity from protein sequence

The high-throughput structure determination pipelines developed by structural genomics programs offer a unique opportunity for data mining. One important question is how protein properties derived from a primary sequence correlate with the protein’s propensity to yield X-ray quality crystals (crystallizability) and 3D X-ray structures. A set of protein properties were computed for over 1,300 proteins that expressed well but were insoluble, and for ~720 unique proteins that resulted in X-ray structures. The correlation of the protein’s iso-electric point and grand average hydropathy (GRAVY) with crystallizability was analyzed for full length and domain constructs of protein targets. In a second step, several additional properties that can be calculated from the protein sequence were added and evaluated. Using statistical analyses we have identified a set of the attributes correlating with a protein’s propensity to crystallize and implemented a Support Vector Machine (SVM) classifier based on these. We have created applications to analyze and provide optimal boundary information for query sequences and to visualize the data. These tools are available via the web site .     

Structural genomics target selection for the New York consortium on membrane protein structure

The New York Consortium on Membrane Protein Structure (NYCOMPS), a part of the Protein Structure Initiative (PSI) in the USA, has as its mission to establish a high-throughput pipeline for determination of novel integral membrane protein structures. Here we describe our current target selection protocol, which applies structural genomics approaches informed by the collective experience of our team of investigators. We first extract all annotated proteins from our reagent genomes, i.e. the 96 fully sequenced prokaryotic genomes from which we clone DNA. We filter this initial pool of sequences and obtain a list of valid targets. NYCOMPS defines valid targets as those that, among other features, have at least two predicted transmembrane helices, no predicted long disordered regions and, except for community nominated targets, no significant sequence similarity in the predicted transmembrane region to any known protein structure. Proteins that feed our experimental pipeline are selected by defining a protein seed and searching the set of all valid targets for proteins that are likely to have a transmembrane region structurally similar to that of the seed. We require sequence similarity aligning at least half of the predicted transmembrane region of seed and target. Seeds are selected according to their feasibility and/or biological interest, and they include both centrally selected targets and community nominated targets. As of December 2008, over 6,000 targets have been selected and are currently being processed by the experimental pipeline. We discuss how our target list may impact structural coverage of the membrane protein space.     

Knowledge-based selection of targets for structural genomics

The problem of rational target selection for protein structure determination in structural genomics projects on microbes is addressed. A flexible computational procedure is described that directly incorporates the whole body of annotation available in the PEDANT genome database into the sequence clustering and selection process in order to identify proteins that are likely to possess currently unknown structural domains. Filtering out gene products based on predicted structural features, such as known three-dimensional structures and transmembrane regions, allows one to reduce the complexity of neighbor relationships between sequences and all but eliminates the need for further partitioning of single-linkage clusters into disjoint protein groups corresponding to homologous families. The results of a large-scale computation experiment in which exemplary target selection for 32 prokaryotic genomes was conducted are presented.     

Structural genomics plucks high-hanging membrane proteins

Recent years have seen the establishment of structural genomics centers that explicitly target integral membrane proteins. Here, we review the advances in targeting these extremely high-hanging fruits of structural biology in high-throughput mode. We observe that the experimental determination of high-resolution structures of integral membrane proteins is increasingly successful both in terms of getting structures and of covering important protein families, for example, from Pfam. Structural genomics has begun to contribute significantly toward this progress. An important component of this contribution is the set up of robotic pipelines that generate a wealth of experimental data for membrane proteins. We argue that prediction methods for the identification of membrane regions and for the comparison of membrane proteins largely suffice to meet the challenges of target selection for structural genomics of membrane proteins. In contrast, we need better methods to prioritize the most promising members in a family of closely related proteins and to annotate protein function from sequence and structure in absence of homology.     

Exploiting a reduced set of weighted average features to improve prediction of DNA-binding residues from 3D structures

Predicting DNA-binding residues from a protein three-dimensional structure is a key task of computational structural proteomics. In the present study, based on machine learning technology, we aim to explore a reduced set of weighted average features for improving prediction of DNA-binding residues on protein surfaces. Via constructing the spatial environment around a DNA-binding residue, a novel weighting factor is first proposed to quantify the distance-dependent contribution of each neighboring residue in determining the location of a binding residue. Then, a weighted average scheme is introduced to represent the surface patch of the considering residue. Finally, the classifier is trained on the reduced set of these weighted average features, consisting of evolutionary profile, interface propensity, betweenness centrality and solvent surface area of side chain. Experimental results on 5-fold cross validation and independent tests indicate that the new feature set are effective to describe DNA-binding residues and our approach has significantly better performance than two previous methods. Furthermore, a brief case study suggests that the weighted average features are powerful for identifying DNA-binding residues and are promising for further study of protein structure-function relationship. The source code and datasets are available upon request.     

Protein surface analysis for function annotation in high-throughput structural genomics pipeline

Structural genomics (SG) initiatives are expanding the universe of protein fold space by rapidly determining structures of proteins that were intentionally selected on the basis of low sequence similarity to proteins of known structure. Often these proteins have no associated biochemical or cellular functions. The SG success has resulted in an accelerated deposition of novel structures. In some cases the structural bioinformatics analysis applied to these novel structures has provided specific functional assignment. However, this approach has also uncovered limitations in the functional analysis of uncharacterized proteins using traditional sequence and backbone structure methodologies. A novel method, named pvSOAR (pocket and void Surface of Amino Acid Residues), of comparing the protein surfaces of geometrically defined pockets and voids was developed. pvSOAR was able to detect previously unrecognized and novel functional relationships between surface features of proteins. In this study, pvSOAR is applied to several structural genomics proteins. We examined the surfaces of YecM, BioH, and RpiB from Escherichia coli as well as the CBS domains from inosine-5'-monosphate dehydrogenase from Streptococcus pyogenes, conserved hypothetical protein Ta549 from Thermoplasm acidophilum, and CBS domain protein mt1622 from Methanobacterium thermoautotrophicum with the goal to infer information about their biochemical function.     

Laboratory scale structural genomics

At Lawrence Livermore National Laboratory, the development of the TB structural genomics consortium crystallization facility has paralleled several local proteomics research efforts that have grown out of gene expression microarray and comparative genomics studies. Collective experience gathered from TB consortium labs and other centers involved in the NIH-NIGMS protein structure initiative allows us to explore the possibilities and challenges of pursuing structural genomics on an academic laboratory scale. We discuss our procedures and protocols for genomic targeting approaches, primer design, cloning, small scale expression screening, scale-up and purification, through to automated crystallization screening and data collection. The procedures are carried out by a small group using a combination of traditional approaches, innovative molecular biochemistry approaches, software automation, and a modest investment in robotic equipment.