Improving the accuracy of high-throughput protein-protein affinity prediction may require better training data

Machine learning methods for protein-protein binding affinity prediction in protein design

Protein-protein interactions govern a wide range of biological activity. A proper estimation of the protein-protein binding affinity is vital to design proteins with high specificity and binding affinity toward a target protein, which has a variety of applications including antibody design in immunotherapy, enzyme engineering for reaction optimization, and construction of biosensors. However, experimental and theoretical modelling methods are time-consuming, hinder the exploration of the entire protein space, and deter the identification of optimal proteins that meet the requirements of practical applications. In recent years, the rapid development in machine learning methods for protein-protein binding affinity prediction has revealed the potential of a paradigm shift in protein design. Here, we review the prediction methods and associated datasets and discuss the requirements and construction methods of binding affinity prediction models for protein design.

A Deep Learning and XGBoost-Based Method for Predicting Protein-Protein Interaction Sites

Knowledge about protein-protein interactions is beneficial in understanding cellular mechanisms. Protein-protein interactions are usually determined according to their protein-protein interaction sites. Due to the limitations of current techniques, it is still a challenging task to detect protein-protein interaction sites. In this article, we presented a method based on deep learning and XGBoost (called DeepPPISP-XGB) for predicting protein-protein interaction sites. The deep learning model served as a feature extractor to remove redundant information from protein sequences. The Extreme Gradient Boosting algorithm was used to construct a classifier for predicting protein-protein interaction sites. The DeepPPISP-XGB achieved the following results: area under the receiver operating characteristic curve of 0.681, a recall of 0.624, and area under the precision-recall curve of 0.339, being competitive with the state-of-the-art methods. We also validated the positive role of global features in predicting protein-protein interaction sites.

RR3DD: an RNA global structure-based RNA three-dimensional structural classification database

The three-dimensional (3D) structure of RNA usually plays an important role in the recognition with RNA-binding protein. Along with the discovering of RNAs, several RNA databases are developed to study the functions of RNA based on sequence, secondary structure, local 3D structural motif and global structure. Based on RNA function and structure, different RNAs are classified and stored in SCOR and DARTS, respectively. The classification of RNA structures is useful in RNA structure prediction and function annotation. However, the SCOR and DARTS are not updated any more. In this study, we present an RNA classification database RR3DD based on RNA fold with the global 3D structural similarity. The RR3DD includes 13,601 RNA chains from PDB and mmCIF format structures which are classified into 780 RNA folds. The RNA chains from PDB and mmCIF format structures are aligned and clustered into 675 and 220 RNA folds, respectively. By analysing the RNA structure in RR3DD, we find that there are 11 clusters with more than 50 members. These clusters include rRNAs, riboswitches, tRNAs and so on. By mapping RR3DD into Rfam, we found that some RNAs without annotation by Rfam can be annotated through structural alignment. For example, we analysed tRNAs and found that tRNA were successfully grouped in RR3DD for which Rfam did not classify them into one family. Finally, we provide a web interface of RR3DD offering functions of browsing RR3DD, annotating RNA 3D structure and finding templates for RNA homology modelling.

Direct molecular evidence for an ancient, conserved developmental toolkit controlling post-transcriptional gene regulation in land plants

In plants, miRNA production is orchestrated by a suite of proteins that control transcription of the pri-miRNA gene, post-transcriptional processing and nuclear export of the mature miRNA. Post-transcriptional processing of miRNAs is controlled by a pair of physically interacting proteins, hyponastic leaves 1 (HYL1) and Dicer-like 1 (DCL1). However, the evolutionary history and structural basis of the HYL1–DCL1 interaction is unknown. Here we use ancestral sequence reconstruction and functional characterization of ancestral HYL1 in vitro and in Arabidopsis thaliana to better understand the origin and evolution of the HYL1–DCL1 interaction and its impact on miRNA production and plant development. We found the ancestral plant HYL1 evolved high affinity for both double-stranded RNA (dsRNA) and its DCL1 partner before the divergence of mosses from seed plants (∼500 Ma), and these high-affinity interactions remained largely conserved throughout plant evolutionary history. Structural modeling and molecular binding experiments suggest that the second of two dsRNA-binding motifs (DSRMs) in HYL1 may interact tightly with the first of two C-terminal DCL1 DSRMs to mediate the HYL1–DCL1 physical interaction necessary for efficient miRNA production. Transgenic expression of the nearly 200 Ma-old ancestral flowering-plant HYL1 in A. thaliana was sufficient to rescue many key aspects of plant development disrupted by HYL1⁻ knockout and restored near-native miRNA production, suggesting that the functional partnership of HYL1–DCL1 originated very early in and was strongly conserved throughout the evolutionary history of terrestrial plants. Overall, our results are consistent with a model in which miRNA-based gene regulation evolved as part of a conserved plant “developmental toolkit.”

PANDA: Predicting the change in proteins binding affinity upon mutations by finding a signal in primary structures

Accurately determining a change in protein binding affinity upon mutations is important to find novel therapeutics and to assist mutagenesis studies. Determination of change in binding affinity upon mutations requires sophisticated, expensive, and time-consuming wet-lab experiments that can be supported with computational methods. Most of the available computational prediction techniques depend upon protein structures that bound their applicability to only protein complexes with recognized 3D structures. In this work, we explore the sequence-based prediction of change in protein binding affinity upon mutation and question the effectiveness of [Formula: see text]-fold cross-validation (CV) across mutations adopted in previous studies to assess the generalization ability of such predictors with no known mutation during training. We have used protein sequence information instead of protein structures along with machine learning techniques to accurately predict the change in protein binding affinity upon mutation. Our proposed sequence-based novel change in protein binding affinity predictor called PANDA performs comparably to the existing methods gauged through an appropriate CV scheme and an external independent test dataset. On an external test dataset, our proposed method gives a maximum Pearson correlation coefficient of 0.52 in comparison to the state-of-the-art existing protein structure-based method called MutaBind which gives a maximum Pearson correlation coefficient of 0.59. Our proposed protein sequence-based method, to predict a change in binding affinity upon mutations, has wide applicability and comparable performance in comparison to existing protein structure-based methods. We made PANDA easily accessible through a cloud-based webserver and python code available at https://sites.google.com/view/wajidarshad/software and https://github.com/wajidarshad/panda, respectively.

High-Throughput Reconstruction of Ancestral Protein Sequence, Structure, and Molecular Function

Ancestral protein sequence reconstruction is a powerful technique for explicitly testing hypotheses about the evolution of molecular function, allowing researchers to meticulously dissect how historical changes in protein sequence impacted functional repertoire by altering the protein's 3D structure. These techniques have provided concrete, experimentally validated insights into ancient evolutionary processes and help illuminate the complex relationship between protein sequence, structure, and function. Inferring the protein family phylogenies on which ancestral sequence reconstruction depends and reconstructing the sequences, themselves, are amenable to high-throughput computational analysis. However, determining the structures of ancestral-reconstructed proteins and characterizing their functions typically rely on time-consuming and expensive laboratory analyses, limiting most current studies to examining a relatively small number of specific hypotheses. For this reason, we have little detailed, unbiased information about how molecular function evolves across large protein family phylogenies. Here we describe a generalized protocol that integrates ancestral sequence reconstruction with structural homology modeling and structure-based molecular affinity prediction to characterize historical changes in protein function across families with thousands of individual sequences. We highlight key steps in the analysis protocol requiring particularly careful attention to avoid introducing potential errors as well as steps for which computationally efficient subroutines can be substituted for more intensive approaches, allowing researchers to scale the analysis up or down, depending on available resources and requirements for reproducibility and scientific rigor. In our view, this approach provides a compelling compliment to more laboratory-intensive procedures, generating important contextual information that can help guide detailed experiments.

Learning protein binding affinity using privileged information

BACKGROUND

Determining protein-protein interactions and their binding affinity are important in understanding cellular biological processes, discovery and design of novel therapeutics, protein engineering, and mutagenesis studies. Due to the time and effort required in wet lab experiments, computational prediction of binding affinity from sequence or structure is an important area of research. Structure-based methods, though more accurate than sequence-based techniques, are limited in their applicability due to limited availability of protein structure data.

RESULTS

In this study, we propose a novel machine learning method for predicting binding affinity that uses protein 3D structure as privileged information at training time while expecting only protein sequence information during testing. Using the method, which is based on the framework of learning using privileged information (LUPI), we have achieved improved performance over corresponding sequence-based binding affinity prediction methods that do not have access to privileged information during training. Our experiments show that with the proposed framework which uses structure only during training, it is possible to achieve classification performance comparable to that which is obtained using structure-based features. Evaluation on an independent test set shows improved performance over the PPA-Pred2 method as well.

CONCLUSIONS

The proposed method outperforms several baseline learners and a state-of-the-art binding affinity predictor not only in cross-validation, but also on an additional validation dataset, demonstrating the utility of the LUPI framework for problems that would benefit from classification using structure-based features. The implementation of LUPI developed for this work is expected to be useful in other areas of bioinformatics as well.

The joint NETTAB/Integrative Bioinformatics 2015 Meeting: aims, topics and outcomes

The 15^th International NETTAB workshop and the 11^th Integrative Bioinformatics Symposium were held together in Bari, on October 14-16, 2016, as Joint NETTAB/IB 2015 Meeting. A special topic for the meeting was "Bioinformatics for ncRNA", but the traditional topics of both meetings series were also included in the event.About 60 scientific contributions were presented, including six keynote lectures, one special guest lecture, and many oral communications and posters. A "Two-Day Hands-on Tutorial" event was organised before the workshop.Selected full papers from some of the best works presented in Bari were submitted either to the Journal of Integrative Bioinformatics or to a purpose Call for a Supplement of BMC Bioinformatics.Here, we provide an overview of meeting aims and scope. We also shortly introduce selected papers that have been either accepted for publication in this Supplement or published in the Journal of Integrative Bioinformatics, for a more complete presentation of the outcomes of the meeting.