Neuropeptides play a variety of roles in many physiological processes and serve as potential therapeutic targets for the treatment some nervous-system disorders. In recent years, there has been tremendous increase number identified neuropeptides. Therefore, we have developed NeuroPep, comprehensive resource neuropeptides, which holds 5949 non-redundant neuropeptide entries originating from 493 organisms belonging to 65 families. neuropeptides invertebrates vertebrates is 3455 2406,...

Neuropeptides are a diverse and complex class of signaling molecules that regulate variety biological processes. provide many opportunities for the discovery new drugs targets treatment wide range diseases, thus, computational tools rapid accurate large-scale identification neuropeptides great significance peptide research drug development. Although several machine learning-based prediction have been developed, there is room improvement in performance interpretability proposed methods. In...

Abstract Motivation: Protein domains are subunits that can fold and evolve independently. Identification of domain boundary locations is often the first step in protein folding function annotations. Most current methods deduce boundaries by sequence-based analysis, which has low accuracy. There no efficient method for predicting discontinuous consist segments from separated sequence regions. As template-based most 3D structure modeling, combining multiple threading alignment information...

We tested two pipelines developed for template-free protein structure prediction in the CASP11 experiment. First, QUARK pipeline constructs models by reassembling fragments of continuously distributed lengths excised from unrelated proteins. Five free-modeling (FM) targets have model successfully constructed with a TM-score above 0.4, including first T0837-D1, which has = 0.736 and RMSD 2.9 Å to native. Detailed analysis showed that success is partly attributed high-resolution contact map...

Abstract The recent emergence of deep learning to characterize complex patterns protein big data reveals its potential address the classic challenges in field mining. Much research has revealed promise as a powerful tool transform into valuable knowledge, leading scientific discoveries and practical solutions. In this review, we summarize publications on predictive approaches mining data. application architectures these methods include multilayer perceptrons, stacked autoencoders, belief...

http://isyslab.info/StraPep.

We report the structure prediction results of a new composite pipeline for template-based modeling (TBM) in 11th CASP experiment. Starting from multiple templates identified by LOMETS based meta-threading programs, QUARK ab initio folding program is extended to generate initial full-length models under strong constraints template alignments. The final atomic are then constructed I-TASSER fragment reassembly simulations, followed fragment-guided molecular dynamic simulation and MQAP-based...

Molecular replacement (MR) is one of the most common techniques used for solving phase problem in X-ray crystal diffraction. The success rate MR however drops quickly when sequence identity between query and templates reduced, while I-TASSER-MR server designed to solve proteins that lack close homologous templates. Starting from a sequence, it first generates full-length models using I-TASSER by iterative structural fragment reassembly. A progressive truncation procedure then editing based...

Abstract Introduction The ocean microbiome represents one of the largest microbiomes and produces nearly half primary energy on planet through photosynthesis or chemosynthesis. Using recent advances in marine genomics, we explore new applications oceanic metagenomes for protein structure function prediction. Results By processing 1.3 TB high-quality reads from Tara Oceans data, obtain 97 million non-redundant genes. Of 5721 Pfam families that lack experimental structures, 2801 have at least...

Accurate delineation of protein domain boundary plays an important role for engineering and structure prediction. Although machine-learning methods are widely used to predict boundary, these approaches often ignore long-range interactions among residues, which have been proven improve the prediction performance. However, how simultaneously model local global further is still a challenging problem.This article employs hybrid deep learning method that combines convolutional neural network gate...

We develop a hierarchical pipeline, ThreaDomEx, for both continuous domain (CD) and discontinuous (DCD) structure predictions. Starting from query sequence, ThreaDomEx first threads it through the PDB to identify multiple templates, where profile of conservation score (DC-score) is derived domain-segment assignment. To further detect DCDs that consist separated segments along boundary-clustering algorithm used refine DCD-linker locations. In case templates do not contain DCDs, assembly...

Transmembrane proteins (TMPs) are essential for cell recognition and communication, they serve as important drug targets in humans. proteins' 3D structures critical determining their functions design but hard to determine even by experimental methods. Although some computational methods have been developed predict transmembrane helices (TMHs) orientation, there is still room improvement. Considering that the pre-trained language model can make full use of massive unlabeled protein sequences...

Abstract We have developed a novel method named AlphaTurn to predict α‐turns in proteins based on the support vector machine (SVM). The prediction was done data set of 469 nonhomologous containing 967 α‐turns. A great improvement performance achieved by using multiple sequence alignment generated PSI‐BLAST as input instead single amino acid sequence. introduction secondary structure information predicted PSIPRED also improved performance. Moreover, we handled very uneven combining cost...

Neuropeptides play critical roles in many biological processes such as growth, learning, memory, metabolism, and neuronal differentiation. A few approaches have been reported for predicting neuropeptides that are cleaved from precursor protein sequences. However, these models cleavage site prediction of precursors were developed using a limited number neuropeptide datasets simple representation models. In addition, universal method sites can be applied to all species is still lacking. this...

Chronic venous insufficiency (CVI) affect a large population, and it cannot heal without doctors' interventions. However, many patients do not get the medical advisory service in time. At same time, doctors also need an assistant tool to classify according severity level of CVI. We propose automatic classification method, named CVI-classifier help patients. In this approach, first, low-level image features are mapped into middle-level semantic by concept classifier, multi-scale model is...

Molecular replacement (MR) often requires templates with high homology to solve the phase problem in X-ray crystallography. I-TASSER-MR has been developed test whether success rate for structure determination of distant-homology proteins could be improved by a combination iterative fragmental structure-assembly simulations progressive sequence truncation designed trim regions variation. The pipeline was tested on two independent protein sets consisting 61 from CASP8 and 100 high-resolution...

Protein domains are the basic units of proteins that can fold, function and evolve independently. domain boundary partition plays an important role in protein structure prediction, understanding their biological functions, annotating evolutionary mechanisms design. Although there many methods have been developed to predict boundaries from sequence over past two decades, is still much room for improvement.In this article, a novel prediction tool called Res-Dom was developed, which based on...

Protein language models have been gaining attention and achieved some exciting breakthroughs in protein modeling tasks compared with the utilization of co-evolutionary biological priors. To overcome shortcomings existing large-scale models, such as high computational complexity large memory consumption, we propose a lightweight model, ProtFlash. It is first model linear based on an strategy, which differs significantly from methods that it primarily utilizes multiple positional encodings...

A variety of protein domain predictors were developed to predict boundaries in recent years, but most them cannot discontinuous domains. Considering nearly 40% multidomain proteins contain one or more domains, we have DomEx enable boundary detect domains by assembling the continuous segments. Discontinuous are predicted matching sequence profile concatenated segments with profiles from a single-domain library derived SCOP and CATH, Pfam. Then matches filtered similarity templates, symmetric...