基于CNN与LSTM模型的蛋白质二级结构预测

蛋白质结构的预测在理解蛋白质结构组成和蛋白质的生物学功能有重要意义,而蛋白质二级结构预测是蛋白质结构预测的重要环节。当PSSM位置特异性进化矩阵被广泛应用于将蛋白质初级结构序列编码作为输入样本后,每个残基可以被表示成二维空间的数据平面,由此文中尝试利用卷积神经网络对其进行训练。文中还设计了另一种卷积神经网络,利用长短记忆网络感知了CNN最后卷积特征面的横向特征和纵向特征后连同卷积神经网络的全连接共同完成分类,最后用ensemble方法对两类卷积神经网络模型进行了整合,最终ensemble方法中包含两类卷积神经网络的六个模型,在CB513蛋白质数据集测得的Q3结果为77.2。     

蛋白质序列的矩阵图谱表达

基于经典HP模型,利用蛋白质序列的矩阵图谱表达法(MGR)及数值刻画的思想提出了一种新的蛋白质序列的比对方法,通过观察蛋白质序列的数值刻画图及计算两蛋白质序列之间的欧氏距离d,对木聚糖酶两家族的蛋白质序列进行了相似性分析.发现被划分为同一木聚糖酶家族的蛋白质序列之间的相似性更大,而且蛋白质序列的相似性程度与分子大小、结构和分子进化相关.     

基于打分矩阵的多类蛋白质折叠子的预测

依据蛋白质折叠子中氨基酸保守性,以氨基酸、氨基酸的极性、氨基酸的电性以及氨基酸的亲—疏水性为参数,从蛋白质的氨基酸序列出发,采用"一对多"的分类策略,通过构建打分矩阵和选取氨基酸序列模式片断,利用5种相似性打分函数对27类折叠子进行识别,最好的预测精度达到83.46%。结果表明,打分矩阵是预测多类蛋白质折叠子有效的方法。     

基于深度学习与多层次信息融合的药物靶标亲和力预测*

药物研发是非常重要但也十分耗费人力物力的过程。利用计算机辅助预测药物与蛋白质亲和力的方法可以极大地加快药物研发过程。药物靶标亲和力预测的关键在于对药物和蛋白质进行准确详细地信息表征。提出一种基于深度学习与多层次信息融合的药物靶标亲和力的预测模型,试图通过综合药物与蛋白质的多层次信息,来获得更好的预测表现。首先将药物表述成分子图和扩展连接指纹两种形式,分别利用图卷积神经网络模块和全连接层进行学习;其次将蛋白质序列和蛋白质K-mer特征分别输入卷积神经网络模块和全连接层来学习蛋白质潜在特征;随后将4个通道学习到的特征进行融合,再利用全连接层进行预测。在两个基准药物靶标亲和力数据集上验证了所提方法的有效性,并与其他已有模型作对比研究。结果说明提出的模型相比基准模型能得到更好的预测性能,表明提出的综合药物与蛋白质多层次信息的药物靶标亲和力预测策略是有效的。     

CBSG-PPI:基于图神经网络的蛋白质-蛋白质相互作用的预测算法

为提高蛋白质-蛋白质相互作用(protein-protein interaction, PPI)预测的准确性,并深入探索细胞信号传导和疾病发生的生物学机制,本文提出一种简称为CBSG-PPI的预测算法。该算法首先利用3层前馈网络来处理蛋白质的k-mer特征,采用CT方法和Bert方法提取蛋白质的氨基酸序列以及使用卷积神经网络提取蛋白质的序列特征,再结合图神经网络和多层感知机来准确预测PPI。与现有的预测技术相比,CBSG-PPI在准确率、 F1分数、召回率和精确率等多个关键性能指标上展现了明显的优势,在公开数据集上分别达到了0.855、 0.853、 0.840和0.866的高分。此外,本算法采用了一种改进的参数调整方法,显著提高了计算效率,其预测速度比传统算法快了约140倍。这一显著的性能提升,不仅证实了CBSG-PPI在预测PPI方面的研究价值,也为未来蛋白质间相互作用网络的构建和分析提供了有用的计算工具。     

支持向量机方法预测蛋白质结构中的二硫键

在蛋白质结构预测的研究中,一个重要的问题就是正确预测二硫键的连接,二硫键的准确预测可以减少蛋白质构像的搜索空间,有利于蛋白质3D结构的预测,本文将预测二硫键的连接问题转化成对连接模式的分类问题,并成功地将支持向量机方法引入到预测工作中。通过对半胱氨酸局域序列连接模式的分类预测,可以由蛋白质的一级结构序列预测该蛋白质的二硫键的连接。结果表明蛋白质的二硫键的连接与半胱氨酸局域序列连接模式有重要联系,应用支持向量机方法对蛋白质结构的二硫键预测取得了良好的结果。     

用离散量的方法识别蛋白质的超二级结构

用离散量的方法,对2208个分辨率在2.5I以上的高精度的蛋白质结构中四类超二级结构进行了识别。从蛋白质一级序列出发,以氨基酸(20种氨基酸加一个空位)和其紧邻关联共同为参数,当序列模式固定长取8个氨基酸残基时,对“822”序列模式3交叉检验的平均预测精度达到78.1%,jack-knife检验的平均预测精度达到76.7%;当序列模式固定长取10个氨基酸残基时,对“1041”序列模式3交叉检验的平均预测精度达到83.1%,jack-knife检验的平均预测精度达到79.8%。     

蛋白质折叠识别方法综述

蛋白质折叠识别算法是蛋白质三维结构预测的重要方法之一,该方法在生物科学的许多方面得到卓有成效的应用。在过去的十年中,我们见证了一系列基于不同计算方式的蛋白质折叠识别方法。在这些计算方法中,机器学习和序列谱-序列谱比对是两种在蛋白质折叠中应用较为广泛和有效的方法。除了计算方法的进展外,不断增大的蛋白质结构数据库也是蛋白质折叠识别的预测精度不断提高的一个重要因素。在这篇文章中,我们将简要地回顾蛋白质折叠中的先进算法。另外,我们也将讨论一些可能可以应用于改进蛋白质折叠算法的策略。     

碳水化合物活性酶数据库( CAZy) 及其研究趋势简

碳水化合物活性酶数据库（ CAZy）是关于能够合成或者分解复杂碳水化合物和糖复合物的酶类的一个数据库资源,其基于蛋白质结构域中的氨基酸序列相似性,将碳水化合物活性酶类归入不同蛋白质家族。 CAZy数据库中包含了碳水化合物酶类的物种来源、酶功能EC分类、基因序列、蛋白质序列及其结构等信息。而随着宏基因组学技术的快速发展,CAZy数据库中家族内序列数据量剧增,这为家族内进一步进行亚家族分类奠定了基础;而蛋白质家族内新一层精细分类的引入可提高亚家族中酶分子功能预测的准确度,进而可指导酶分子理性设计来提高特定功能酶组分设计的成功概率,从而推动生物质转化产业的发展。     

碳水化合物活性酶数据库(CAZy)及其研究趋势

碳水化合物活性酶数据库(CAZy)是关于能够合成或者分解复杂碳水化合物和糖复合物的酶类的一个数据库资源,其基于蛋白质结构域中的氨基酸序列相似性,将碳水化合物活性酶类归入不同蛋白质家族。CAZy数据库中包含了碳水化合物酶类的物种来源、酶功能EC分类、基因序列、蛋白质序列及其结构等信息。而随着宏基因组学技术的快速发展,CAZy数据库中家族内序列数据量剧增,这为家族内进一步进行亚家族分类奠定了基础;而蛋白质家族内新一层精细分类的引入可提高亚家族中酶分子功能预测的准确度,进而可指导酶分子理性设计来提高特定功能酶组分设计的成功概率,从而推动生物质转化产业的发展。     

Variations on probabilistic suffix trees: statistical modeling and prediction of protein families

MOTIVATION: We present a method for modeling protein families by means of probabilistic suffix trees (PSTs). The method is based on identifying significant patterns in a set of related protein sequences. The patterns can be of arbitrary length, and the input sequences do not need to be aligned, nor is delineation of domain boundaries required. The method is automatic, and can be applied, without assuming any preliminary biological information, with surprising success. Basic biological considerations such as amino acid background probabilities, and amino acids substitution probabilities can be incorporated to improve performance. RESULTS: The PST can serve as a predictive tool for protein sequence classification, and for detecting conserved patterns (possibly functionally or structurally important) within protein sequences. The method was tested on the Pfam database of protein families with more than satisfactory performance. Exhaustive evaluations show that the PST model detects much more related sequences than pairwise methods such as Gapped-BLAST, and is almost as sensitive as a hidden Markov model that is trained from a multiple alignment of the input sequences, while being much faster.     

Challenges in homology search: HMMER3 and convergent evolution of coiled-coil regions

Detection of protein homology via sequence similarity has important applications in biology, from protein structure and function prediction to reconstruction of phylogenies. Although current methods for aligning protein sequences are powerful, challenges remain, including problems with homologous overextension of alignments and with regions under convergent evolution. Here, we test the ability of the profile hidden Markov model method HMMER3 to correctly assign homologous sequences to >13 000 manually curated families from the Pfam database. We identify problem families using protein regions that match two or more Pfam families not currently annotated as related in Pfam. We find that HMMER3 E-value estimates seem to be less accurate for families that feature periodic patterns of compositional bias, such as the ones typically observed in coiled-coils. These results support the continued use of manually curated inclusion thresholds in the Pfam database, especially on the subset of families that have been identified as problematic in experiments such as these. They also highlight the need for developing new methods that can correct for this particular type of compositional bias.     

Topological maps of protein sequences

A new method based on neural networks to cluster proteins into families is described. The network is trained with the Kohonen unsupervised learning algorithm, using matrix pattern representations of the protein sequences as inputs. The components (x, y) of these 20×20 matrix patterns are the normalized frequencies of all pairs xy of amino acids in each sequence. We investigate the influence of different learning parameters in the final topological maps obtained with a learning set of ten proteins belonging to three established families. In all cases, except in those where the synaptic vectors remains nearly unchanged during learning, the ten proteins are correctly classified into the expected families. The classification by the trained network of mutated or incomplete sequences of the learned proteins is also analysed. The neural network gives a correct classification for a sequence mutated in 21.5%±7% of its amino acids and for fragments representing 7.5%±3% of the original sequence. Similar results were obtained with a learning set of 32 proteins belonging to 15 families. These results show that a neural network can be trained following the Kohonen algorithm to obtain topological maps of protein sequences, where related proteins are finally associated to the same winner neuron or to neighboring ones, and that the trained network can be applied to rapidly classify new sequences. This approach opens new possibilities to find rapid and efficient algorithms to organize and search for homologies in the whole protein database.     

Classification of protein families and detection of the determinant residues with an improved self-organizing map

 Using a SOM (self-organizing map) we can classify sequences within a protein family into subgroups that generally correspond to biological subcategories. These maps tend to show sequence similarity as proximity in the map. Combining maps generated at different levels of resolution, the structure of relations in protein families can be captured that could not otherwise be represented in a single map. The underlying representation of maps enables us to retrieve characteristic sequence patterns for individual subgroups of sequences. Such patterns tend to correspond to functionally important regions. We present a modified SOM algorithm that includes a convergence test that dynamically controls the learning parameters to adapt them to the learning set instead of being fixed and externally optimized by trial and error. Given the variability of protein family size and distribution, the addition of this feature is necessary. The method is successfully tested with a number of families. The rab family of small GTPases is used to illustrate the performance of the method. Received: 25 July 1996 / Accepted in revised form: 13 February 1997     

Identification of functionally conserved residues with the use of entropy-variability plots

We introduce sequence entropy-variability plots as a method of analyzing families of protein sequences, and demonstrate this for three well-known sequence families: globins, ras-like proteins, and serine-proteases. The location of an aligned residue position in the entropy-variability plot correlates with structural characteristics, and with known facts about the roles of individual amino acids in the function of these proteins. The large numbers of known sequences in these families allowed us to introduce new filtering methods for variability patterns. The results are discussed in terms of a simple evolutionary model for functional proteins.     

Computational neurogenetic modelling: a pathway to new discoveries in genetic neuroscience

The paper presents a methodology for using computational neurogenetic modelling (CNGM) to bring new original insights into how genes influence the dynamics of brain neural networks. CNGM is a novel computational approach to brain neural network modelling that integrates dynamic gene networks with artificial neural network model (ANN). Interaction of genes in neurons affects the dynamics of the whole ANN model through neuronal parameters, which are no longer constant but change as a function of gene expression. Through optimization of interactions within the internal gene regulatory network (GRN), initial gene/protein expression values and ANN parameters, particular target states of the neural network behaviour can be achieved, and statistics about gene interactions can be extracted. In such a way, we have obtained an abstract GRN that contains predictions about particular gene interactions in neurons for subunit genes of AMPA, GABAA and NMDA neuro-receptors. The extent of sequence conservation for 20 subunit proteins of all these receptors was analysed using standard bioinformatics multiple alignment procedures. We have observed abundance of conserved residues but the most interesting observation has been the consistent conservation of phenylalanine (F at position 269) and leucine (L at position 353) in all 20 proteins with no mutations. We hypothesise that these regions can be the basis for mutual interactions. Existing knowledge on evolutionary linkage of their protein families and analysis at molecular level indicate that the expression of these individual subunits should be coordinated, which provides the biological justification for our optimized GRN.     

Codep: maximizing co-evolutionary interdependencies to discover interacting proteins

Approaches for the determination of interacting partners from different protein families (such as ligands and their receptors) have made use of the property that interacting proteins follow similar patterns and relative rates of evolution. Interacting protein partners can then be predicted from the similarity of their phylogenetic trees or evolutionary distances matrices. We present a novel method called Codep, for the determination of interacting protein partners by maximizing co-evolutionary signals. The order of sequences in the multiple sequence alignments from two protein families is determined in such a manner as to maximize the similarity of substitution patterns at amino acid sites in the two alignments and, thus, phylogenetic congruency. This is achieved by maximizing the total number of interdependencies of amino acids sites between the alignments. Once ordered, the corresponding sequences in the two alignments indicate the predicted interacting partners. We demonstrate the efficacy of this approach with computer simulations and in analyses of several protein families. A program implementing our method, Codep, is freely available to academic users from our website: http://www.uhnresearch.ca/labs/tillier/.     

Identification of members of gene families in Arabidopsis thaliana by contig construction from partial cDNA sequences: 106 genes encoding 50 cytoplasmic ribosomal proteins

Partial cDNA sequencing to obtain expressed sequence tags (ESTs) has led to the identification of tags to about 8000 of the estimated 20 000 genes in Arabidopsis thaliana . This figure represents four to five times the number of complete coding sequences from this organism available in international databases. In contrast to mammals, many proteins are encoded by multigene families in A. thaliana . Using ribosomal protein gene families as an example, it is possible to construct relatively long sequences from overlapping ESTs which are of sufficiently high quality to be able to unambiguously identify tags to individual members of multigene families, even when the sequences are highly conserved. A total of 106 genes encoding 50 different cytoplasmic ribosomal protein types have been identified, most proteins being encoded by at least two and up to four genes. Coding sequences of members of individual gene families are almost always very highly conserved and derived amino acid sequences are almost, if not completely, identical in the vast majority of cases. Sequence divergence is observed in untranslated regions which allows the definition of gene-specific probes. The method can be used to construct high-quality tags to any protein.     

UPSEC: an algorithm for classifying unaligned protein sequences into functional families.

To classify proteins into functional families based on their primary sequences, popular algorithms such as the k-NN-, HMM-, and SVM-based algorithms are often used. For many of these algorithms to perform their tasks, protein sequences need to be properly aligned first. Since the alignment process can be error-prone, protein classification may not be performed very accurately. To improve classification accuracy, we propose an algorithm, called the Unaligned Protein SEquence Classifier (UPSEC), which can perform its tasks without sequence alignment. UPSEC makes use of a probabilistic measure to identify residues that are useful for classification in both positive and negative training samples, and can handle multi-class classification with a single classifier and a single pass through the training data. UPSEC has been tested with real protein data sets. Experimental results show that UPSEC can effectively classify unaligned protein sequences into their corresponding functional families, and the patterns it discovers during the training process can be biologically meaningful.     

Identification of novel families of membrane proteins from the model plant Arabidopsis thaliana

MOTIVATION: The completion of the Arabidopsis genome offers the first opportunity to analyze all of the membrane protein sequences of a plant. The majority of integral membrane proteins including transporters, channels, and pumps contain hydrophobic alpha-helices and can be selected based on TransMembrane Spanning (TMS) domain prediction. By clustering the predicted membrane proteins based on sequence, it is possible to sort the membrane proteins into families of known function, based on experimental evidence or homology, or unknown function. This provides a way to identify target sequences for future functional analysis. RESULTS: An automated approach was used to select potential membrane protein sequences from the set of all predicted proteins and cluster the sequences into related families. The recently completed sequence of Arabidopsis thaliana, a model plant, was analyzed. Of the 25,470 predicted protein sequences 4589 (18%) were identified as containing two or more membrane spanning domains. The membrane protein sequences clustered into 628 distinct families containing 3208 sequences. Of these, 211 families (1764 sequences) either contained proteins of known function or showed homology to proteins of known function in other species. However, 417 families (1444 sequences) contained only sequences with no known function and no homology to proteins of known function. In addition, 1381 sequences did not cluster with any family and no function could be assigned to 1337 of these.