Department of Biotechnology, Bhupat and Jyoti Mehta School of BioSciences, Indian Institute of Technology Madras, Chennai 600 036, Tamilnadu, India.
Comb Chem High Throughput Screen. 2023;26(4):769-777. doi: 10.2174/1386207325666220520102316.
Alzheimer's disease (AD) is the most common neurodegenerative disorder that affects the neuronal system and leads to memory loss. Many coding gene variants are associated with this disease and it is important to characterize their annotations.
We collected the Alzheimer's disease-causing and neutral mutations from different databases. For each mutation, we computed the different features from protein sequence. Further, these features were used to build a Bayes network-based machine-learning algorithm to discriminate between the disease-causing and neutral mutations in AD.
We have constructed a comprehensive dataset of 314 Alzheimer's disease-causing and 370 neutral mutations and explored their characteristic features such as conservation scores, positionspecific scoring matrix (PSSM) profile, and the change in hydrophobicity, different amino acid residue substitution matrices and neighboring residue information for identifying the disease-causing mutations. Utilizing these features, we have developed a disease-specific tool named Alz-disc, for discriminating the disease-causing and neutral mutations using sequence information alone. The performance of the present method showed an accuracy of 89% for independent test set, which is 13% higher than available generic methods. This method is freely available as a web server at https://web.iitm.ac.in/bioinfo2/alzdisc/.
This study is useful to annotate the effect of new variants and develop mutation specific drug design strategies for Alzheimer's disease.
阿尔茨海默病（AD）是最常见的神经退行性疾病，影响神经元系统，导致记忆丧失。许多编码基因变异与这种疾病有关，因此重要的是对其注释进行特征描述。
我们从不同的数据库中收集了导致阿尔茨海默病的突变和中性突变。对于每个突变，我们从蛋白质序列中计算了不同的特征。此外，这些特征被用于构建基于贝叶斯网络的机器学习算法，以区分 AD 中的致病突变和中性突变。
我们构建了一个包含 314 个阿尔茨海默病致病突变和 370 个中性突变的综合数据集，并探索了它们的特征，如保守评分、位置特异性评分矩阵（PSSM）谱以及疏水性变化、不同的氨基酸残基取代矩阵和相邻残基信息，以识别致病突变。利用这些特征，我们开发了一种名为 Alz-disc 的疾病特异性工具，用于仅使用序列信息区分致病突变和中性突变。该方法在独立测试集上的准确率为 89%，比现有的通用方法高 13%。该方法可作为一个免费的网络服务器，网址为 https://web.iitm.ac.in/bioinfo2/alzdisc/。
这项研究有助于注释新变体的影响，并为阿尔茨海默病开发突变特异性药物设计策略。
