TP-MV: Therapeutic protein prediction by multi-view learning

Background: Therapeutic peptide prediction is critical for drug development and therapy. Researchers have been studying this essential task, developing several computational methods to identify different therapeutic peptide types. Objective: Most predictors are the specific methods for certain peptides. Currently, developing methods to predict the presence of multiple peptides remains a challenging problem. Moreover, it is still challenging to combine different features to make the therapeutic prediction. Method: In this paper, we proposed a new ensemble method TP-MV for general therapeutic peptide recognition. TP-MV is developed using the stacking framework in conjunction with the KNN, SVM, ET, RF, and XGB. Then TP-MV constructs a multi-view learning model as meta-classifiers to extract the discriminative feature for different peptides. Results: In the experiment, the proposed method outperforms the other existing methods on the benchmark datasets, indicating that the proposed method has the ability to predict multiple therapeutic peptides simultaneously. Conclusion: The TP-MV is a useful tool for predicting therapeutic peptides.

Cell-Penetrating Peptides Predicted From CASC3, AKIP1, and AHRR Proteins

Peptides can be used as research tools and for diagnostic or therapeutic applications. Peptides, alongside small molecules and antibodies, are used and are gaining further interest as protein-protein interaction (PPI) modulators. Peptides have high target specificity and high affinity, but, unlike small molecule modulators, they are not able to cross the cell membranes to reach their intracellular targets. To overcome this limitation, the special property of the cell-penetrating peptides (CPPs) could benefit their cause. CPPs are a class of peptides that can enter the cells and with them also deliver the attached cargoes. Today, with the advancement of in silico prediction tools and the availability of protein databases, designing new and multifunctional peptides that are able to reach intracellular targets and inhibit certain cellular processes in a very specific manner is reachable. Although there are several efficient CPP sequences already known, the discovery of new CPPs is crucial for the development of efficient delivery methods for both biotechnological and therapeutic applications. In this work, we chose 10 human nuclear proteins from which we predicted new potential CPP sequences by using three different CPP predictors: cell-penetrating peptide prediction tool, CellPPD, and SkipCPP-Pred. From each protein, one predicted CPP sequence was synthesized and its internalization into cells was assessed. Out of the tested sequences, three peptides displayed features characteristic to CPPs. These peptides and also the predicted peptide sequences could be used to design and modify new CPPs. In this work, we show that we can use protein sequences as input for generating new peptides with cell internalization properties. Three new CPPs, AHRR8-24, CASC3251-264, and AKIP127-37, can be further used for the delivery of other cargoes or designed into multifunctional peptides with capability of internalizing cells.

In Silico Bioactive Peptide Prediction from The Enzymatic Hydrolysates of Edible Seaweed Rubisco Large Chain

Seaweeds are one of the ancient food supplements on Earth. Especially Asian countries use seaweeds as the fundamental ingredient in their cuisine. Seaweeds are photosynthetic organisms living in aquatic ecosystems and in the coastal territories. Seaweeds out of farm areas are frequently observed as coastal wastes. However, seaweeds are outstanding sources for bioactive substances and investigation bioactive properties of seaweed RuBisCO has never been done. RuBisCO is the most abundant protein on Earth but a vast amount of RuBisCO goes through waste. In this study, bioactive peptide prediction of frequently consumed seaweed RuBisCO proteins were analyzed in silico to identify possible bioactive peptides as substitute or support for grain, meat, and dairy based bioactive peptides. A huge portion of peptides were di-, tri- peptides with IC50 values less than 300 µM according to the comparison of BIOPEP database. Including gastric digestion, more than half of the peptides showed DDP-IV and ACE inhibitory activity followed by antioxidant properties. Also, novel antiinflammatory and anti-cancer peptides were found through in silico analysis.

EffectorP 3.0: prediction of apoplastic and cytoplasmic effectors in fungi and oomycetes

Many fungi and oomycete species are devasting plant pathogens. These eukaryotic filamentous pathogens secrete effector proteins to facilitate plant infection. Fungi and oomycete pathogens have diverse infection strategies and their effectors generally do not share sequence homology. However, they occupy similar host environments, either the plant apoplast or plant cytoplasm, and may therefore share some unifying properties based on the requirements of these host compartments. Here we exploit these biological signals and present the first classifier (EffectorP 3.0) that uses two machine learning models: one trained on apoplastic effectors and one trained on cytoplasmic effectors. EffectorP 3.0 accurately predicts known apoplastic and cytoplasmic effectors in fungal and oomycete secretomes with low estimated false positive rates of 3% and 8%, respectively. Cytoplasmic effectors have a higher proportion of positively charged amino acids, whereas apoplastic effectors are enriched for cysteine residues. The combination of fungal and oomycete effectors in training leads to a higher number of predicted cytoplasmic effectors in biotrophic fungi. EffectorP 3.0 expands predicted effector repertoires beyond small, cysteine-rich secreted proteins in fungi and RxLR-motif containing secreted proteins in oomycetes. We show that signal peptide prediction is essential for accurate effector prediction, as EffectorP 3.0 recognizes a cytoplasmic signal also in intracellular, non-secreted proteins. EffectorP 3.0 is available at http://effectorp.csiro.au.

SignalP 6.0 achieves signal peptide prediction across all types using protein language models

Signal peptides (SPs) are short amino acid sequences that control protein secretion and translocation in all living organisms. As experimental characterization of SPs is costly, prediction algorithms are applied to predict them from sequence data. However, existing methods are unable to detect all known types of SPs. We introduce SignalP 6.0, the first model capable of detecting all five SP types. Additionally, the model accurately identifies the positions of regions within SPs, revealing the defining biochemical properties that underlie the function of SPs in vivo. Results show that SignalP 6.0 has improved prediction performance, and is the first model to be applicable to metagenomic data. SignalP 6.0 is available at https://services.healthtech.dtu.dk/service.php?SignalP-6.0

Epitope-Based Peptide prediction of vaccine against European bat lyssaviruses type 2 Glycoprotein G with Immunoinformatic Approach

Objective: European bat lyssaviruses (EBLV) type 2 is present in many European countries. Infection is usually seen in bats, the primary reservoirs of the viruses. Human deaths have been documented within few days following bat exposures. So, it is very useful to design an insilco peptide vaccine for European bat lyssaviruses type 2 virus using glycoprotein G as an immunogen to stimulate protective immune response. Results: B cell tests were conducted for Bepipred with 15 conserved epitopes, Emini surface accessibility prediction with 7 conserved epitopes in the surface and Kolaskar and Tongaonkar antigenicity tested with three conserved epitopes being antigenic. 357 conserved epitopes were predicted to interact with different MHC-1 alleles with (IC50) ≤500 while 282 conserved epitopes found to interact with MHC-II alleles with IC50≤ 1000. Among all tested epitopes for world population coverage the epitope VFSYMELKV binding to MHC11 alleles was 97.94% and it found to bind 10 different alleles that indicate strong potential to formulate peptide vaccine for lyssaviruses type 2 virus. To the best of our knowledge this is the first study to propose peptide vaccine for European bat lyssavirus type 2.

In vitro proteasome processing of neo-splicetopes does not predict their presentation in vivo

Proteasome catalyzed peptide splicing (PCPS) of cancer-driving antigens could generate attractive neoepitopes to be targeted by TCR-based adoptive T cell therapy. Based on a spliced peptide prediction algorithm TCRs were generated against putative KRASG12V and RAC2P29L derived neo-splicetopes with high HLA-A*02:01 binding affinity. TCRs generated in mice with a diverse human TCR repertoire specifically recognized the respective target peptides with high efficacy. However, we failed to detect any neo-splicetope specific T cell response when testing the in vivo neo-splicetope generation and obtained no experimental evidence that the putative KRASG12V- and RAC2P29L-derived neo-splicetopes were naturally processed and presented. Furthermore, only the putative RAC2P29L-derived neo-splicetopes was generated by in vitro PCPS. The experiments pose severe questions on the notion that available algorithms or the in vitro PCPS reaction reliably simulate in vivo splicing and argue against the general applicability of an algorithm-driven 'reverse immunology' pipeline for the identification of cancer-specific neo-splicetopes.

Comparison of Current Methods for Signal Peptide Prediction in Phytoplasmas

Although phytoplasma studies are still hampered by the lack of axenic cultivation methods, the availability of genome sequences allowed dramatic advances in the characterization of the virulence mechanisms deployed by phytoplasmas, and highlighted the detection of signal peptides as a crucial step to identify effectors secreted by phytoplasmas. However, various signal peptide prediction methods have been used to mine phytoplasma genomes, and no general evaluation of these methods is available so far for phytoplasma sequences. In this work, we compared the prediction performance of SignalP versions 3.0, 4.0, 4.1, 5.0 and Phobius on several sequence datasets originating from all deposited phytoplasma sequences. SignalP 4.1 with specific parameters showed the most exhaustive and consistent prediction ability. However, the configuration of SignalP 4.1 for increased sensitivity induced a much higher rate of false positives on transmembrane domains located at N-terminus. Moreover, sensitive signal peptide predictions could similarly be achieved by the transmembrane domain prediction ability of TMHMM and Phobius, due to the relatedness between signal peptides and transmembrane regions. Beyond the results presented herein, the datasets assembled in this study form a valuable benchmark to compare and evaluate signal peptide predictors in a field where experimental evidence of secretion is scarce. Additionally, this study illustrates the utility of comparative genomics to strengthen confidence in bioinformatic predictions.

Overall performance and taxonomic bias of antimicrobial peptide prediction tools in invertebrates

Invertebrate antimicrobial peptides (AMPs) are at the forefront in the search for agents of therapeutic utility against multi-resistant microbial pathogens, and in recent years substantial advances took place in the in silico prediction of antimicrobial function of amino acid sequences. A yet neglected aspect is taxonomic bias in the performance of these tools. Owing to differences in the prediction algorithms and used training data sets between tools, and phylogenetic differences in sequence diversity, physicochemical properties and evolved biological functions of AMPs between taxa, notable discrepancies may exist in performance between the currently available prediction tools. Here, we tested if there is a taxonomic bias in prediction power in 8 tools with a total of 15 prediction algorithms, in 19 invertebrate taxa. We found that most of the tools exhibited considerable variation in performance between tested invertebrate groups, based on which we provide guidance in choosing the adequate prediction tool for specific taxa.