A class of secondary structure prediction algorithms use the information from the statistics of the residue pairs found in secondary structural elements. Because the protein folding process is dominated by backbone hydrogen bonding, an approach based on backbone hydrogen-bonded residue pairings would improve the predicting capabilities of these class algorithms. The reliability of the prediction algorithms depends on the quality of the statistics, therefore, of the data set. In this study, it was aimed to determine the propensities of the backbone hydrogen-bonded residue pairings for secondary structural elements of α-helix and β-sheet in globular proteins using a new and comprehensive data set created from the peptides deposited in Worldwide Protein Data Bank. A master data set including 4882 globular peptide chains with resolution better than 2.5 Å, sequence identity smaller than 25% and length of no shorter than 100 residues were created. Separate data sub sets also were created for helix and sheet structures from master set and each sub set includes 4594 and 4483 chains, respectively. Backbone hydrogen-bonded residue pairings in helices and sheets were detected and the propensities of them were represented as odds ratios (observed/[random or expected]) in matrices. Propensities assigned by this study to the residue pairings in secondary structural elements (as helix, overall strands, parallel strands and antiparallel strands) differ from the previous studies by 19 to 34%. These dissimilarities are important and they would cause further improvements in secondary structure prediction algorithms.

中文翻译：

---

球状蛋白二级结构中氨基酸配对的倾向。

一类二级结构预测算法使用来自二级结构元素中残基对的统计信息。由于蛋白质折叠过程由主链氢键主导，因此基于主链氢键残基配对的方法将提高这些类算法的预测能力。预测算法的可靠性取决于统计数据的质量，因此取决于数据集的质量。在这项研究中，目的是使用全球蛋白质数据中保存的肽段建立的新的综合数据集，确定球形蛋白质中α-螺旋和β-sheet二级结构元素的骨架氢键残基配对的倾向。银行。主数据集包括4882个分辨率高于2的球状肽链。5Å，序列同一性小于25％，长度不小于100个残基。还为螺旋结构和图纸结构从主集中创建了单独的数据子集，每个子​​集分别包括4594和4483链。检测到螺旋和薄片中的主链氢键残基配对，并以矩阵中的优势比（观察/ [随机或预期]）表示它们的倾向。这项研究分配给二级结构元素（如螺旋，整体链，平行链和反平行链）中残基配对的倾向与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。还为螺旋结构和图纸结构从主集中创建了单独的数据子集，每个子​​集分别包括4594和4483链。检测到螺旋和薄片中的主链氢键残基配对，并以矩阵中的优势比（观察/ [随机或预期]）表示它们的倾向。这项研究分配给二级结构元素（如螺旋，整体链，平行链和反平行链）中残基配对的倾向与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。还为螺旋结构和图纸结构从主集中创建了单独的数据子集，每个子​​集分别包括4594和4483链。检测到螺旋和薄片中的主链氢键残基配对，并以矩阵中的优势比（观察/ [随机或预期]）表示它们的倾向。这项研究分配给二级结构元素（如螺旋，整体链，平行链和反平行链）中残基配对的倾向与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。检测到螺旋和薄片中的主链氢键残基配对，并以矩阵中的优势比（观察/ [随机或预期]）表示它们的倾向。这项研究分配给二级结构元素（如螺旋，整体链，平行链和反平行链）中残基配对的倾向与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。检测到螺旋和薄片中的主链氢键残基配对，并以矩阵中的优势比（观察/ [随机或预期]）表示它们的倾向。这项研究分配给二级结构元素（如螺旋，整体链，平行链和反平行链）中残基配对的倾向与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。平行链和反平行链）与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。平行链和反平行链）与以前的研究相差19％至34％。这些差异很重要，它们将导致二级结构预测算法的进一步改进。