BACKGROUND: Innovations in protein engineering offer promising solutions for redesigning allergenic proteins to minimize adverse reactions in sensitive individuals. Earlier models for predicting allergenicity have relied on the knowledge of physicochemical properties and sequence homology to assess the potential risk. However, to better understand the allergenic proteins’ sequence features, we need a novel sequence-based deep learning model for predicting allergenicity. RESULTS: We present a novel AI-based tool, AllergenAI, to quantify the allergenic potential of a protein’s sequence without using any other known features. Our study utilized allergenic protein sequence data archived in the three well-established databases, SDAP 2.0, COMPARE, and AlgPred 2, to train a convolutional neural network and assessed its prediction performance by cross-validation. We then used AllergenAI to find novel potential proteins of the cupin family in date palm, spinach, maize, and red clover plants with a high allergenicity score that might have an adverse allergenic effect on sensitive individuals. By analyzing the feature importance scores (FIS) of vicilins, we identified a proline-alanine-rich (P-A) motif in the top 50% of FIS regions that overlapped with known IgE epitope regions of vicilin allergens. We then used the approximately 1600 allergen structures in our SDAP database, in a pilot study to show the potential to incorporate 3D information in a CNN model. The prediction quality was slightly increased. CONCLUSION: Our allergenicity prediction study through the development of AllergenAI provides a foundation for identifying the critical features that distinguish allergenic proteins.

The phenomenon of cross-reactivity between allergenic proteins plays an important role to understand how the immune system recognizes different antigen proteins. Allergen proteins are known to cross-react if their sequence comparison shows a high sequence identity which also implies that the proteins have a similar 3D fold. In such cases, linear sequence alignment methods are frequently used to predict cross-reactivity between allergenic proteins. However, the prediction of cross-reactivity between distantly related allergens continues to be a challenging task. To overcome this problem, we developed a new structure-based computational method, Cross-React, to predict cross-reactivity between allergenic proteins available in the Structural Database of Allergens (SDAP). Our method is based on the hypothesis that we can find surface patches on 3D structures of potential allergens with amino acid compositions similar to an epitope in a known allergen. We applied the Cross-React method to a diverse set of seven allergens, and successfully identified several cross-reactive allergens with high to moderate sequence identity which have also been experimentally shown to cross-react. Based on these findings, we suggest that Cross-React can be used as a predictive tool to assess protein allergenicity and cross-reactivity. Availability and Implementation: : Cross-React is available at: http://curie.utmb.edu/Cross-React.html. Contact: [email protected].

UNLABELLED: A new web server, InterProSurf, predicts interacting amino acid residues in proteins that are most likely to interact with other proteins, given the 3D structures of subunits of a protein complex. The prediction method is based on solvent accessible surface area of residues in the isolated subunits, a propensity scale for interface residues and a clustering algorithm to identify surface regions with residues of high interface propensities. Here we illustrate the application of InterProSurf to determine which areas of Bacillus anthracis toxins and measles virus hemagglutinin protein interact with their respective cell surface receptors. The computationally predicted regions overlap with those regions previously identified as interface regions by sequence analysis and mutagenesis experiments. AVAILABILITY: The InterProSurf web server is available at http://curie.utmb.edu/

MOTIVATION: Identification of short conserved sequence motifs common to a protein family or superfamily can be more useful than overall sequence similarity in suggesting the function of novel gene products. Locating motifs still requires expert knowledge, as automated methods using stringent criteria may not differentiate subtle similarities from statistical noise. RESULTS: We have developed a novel automatic method, based on patterns of conservation of 237 physical-chemical properties of amino acids in aligned protein sequences, to find related motifs in proteins with little or no overall sequence similarity. As an application, our web-server MASIA identified 12 property-based motifs in the apurinic/apyrimidinic endonuclease (APE) family of DNA-repair enzymes of the DNase-I superfamily. Searching with these motifs located distantly related representatives of the DNase-I superfamily, such as Inositol 5'-polyphosphate phosphatases in the ASTRAL40 database, using a Bayesian scoring function. Other proteins containing APE motifs had no overall sequence or structural similarity. However, all were phosphatases and/or had a metal ion binding active site. Thus our automated method can identify discrete elements in distantly related proteins that define local structure and aspects of function. We anticipate that our method will complement existing ones to functionally annotate novel protein sequences from genomic projects. AVAILABILITY: MASIA WEB site: http://www.scsb.utmb.edu/masia/masia.html SUPPLEMENTARY INFORMATION: The dendrogram of 42 APE sequences used to derive motifs is available on http://www.scsb.utmb.edu/comp_biol.html/DNA_repair/publication.html

MOTIVATION: Many sequences, and in some cases structures, of proteins that induce an allergic response in atopic individuals have been determined in recent years. This data indicates that allergens, regardless of source, fall into discreet protein families. Similarities in the sequence may explain clinically observed cross-reactivities between different biological triggers. However, previously available allergy databases group allergens according to their biological sources, or observed clinical cross-reactivities, without providing data about the proteins. A computer-aided data mining system is needed to compare the sequential and structural details of known allergens. This information will aid in predicting allergenic cross-responses and eventually in determining possible common characteristics of IgE recognition. RESULTS: The new web-based Structural Database of Allergenic Proteins (SDAP) permits the user to quickly compare the sequence and structure of allergenic proteins. Data from literature sources and previously existing lists of allergens are combined in a MySQL interactive database with a wide selection of bioinformatics applications. SDAP can be used to rapidly determine the relationship between allergens and to screen novel proteins for the presence of IgE or T-cell epitopes they may share with known allergens. Further, our novel similarity search method, based on five dimensional descriptors of amino acid properties, can be used to scan the SDAP entries with a peptide sequence. For example, when a known IgE binding epitope from shrimp tropomyosin was used as a query, the method rapidly identified a similar sequence in known shellfish and insect allergens. This prediction of cross-reactivity between allergens is consistent with clinical observations. AVAILABILITY: SDAP is available on the web at http://fermi.utmb.edu/SDAP/index.html

BACKGROUND: Total sequence decomposition, using the web-based MASIA tool, identifies areas of conservation in aligned protein sequences. By structurally annotating these motifs, the sequence can be parsed into individual building blocks, molecular legos ("molegos"), that can eventually be related to function. Here, the approach is applied to the apurinic/apyrimidinic endonuclease (APE) DNA repair proteins, essential enzymes that have been highly conserved throughout evolution. The APEs, DNase-1 and inositol 5'-polyphosphate phosphatases (IPP) form a superfamily that catalyze metal ion based phosphorolysis, but recognize different substrates. RESULTS: MASIA decomposition of APE yielded 12 sequence motifs, 10 of which are also structurally conserved within the family and are designated as molegos. The 12 motifs include all the residues known to be essential for DNA cleavage by APE. Five of these molegos are sequentially and structurally conserved in DNase-1 and the IPP family. Correcting the sequence alignment to match the residues at the ends of two of the molegos that are absolutely conserved in each of the three families greatly improved the local structural alignment of APEs, DNase-1 and synaptojanin. Comparing substrate/product binding of molegos common to DNase-1 showed that those distinctive for APEs are not directly involved in cleavage, but establish protein-DNA interactions 3' to the abasic site. These additional bonds enhance both specific binding to damaged DNA and the processivity of APE1. CONCLUSION: A modular approach can improve structurally predictive alignments of homologous proteins with low sequence identity and reveal residues peripheral to the traditional "active site" that control the specificity of enzymatic activity.

SUMMARY: MASIA is a software tool for pattern recognition in multiple aligned protein sequences. MASIA converts a sequence to a properties matrix that can be scanned in both vertical and horizontal steps. Consistent patterns are recognized based on the statistical significance of their occurrence. Preset macros can be altered on-line to seek any combination of amino acid properties or sequence characteristics. MASIA output can be used directly by our programs to predict the 3D structure of proteins. AVAILABILITY: Access MASIA at http://www.scsb.utmb.edu/masia/ma sia.html.