Computational Molecular Biology 2024, Vol.14, No.1 http://bioscipublisher.com/index.php/cmb © 2024 BioSci Publisher, registered at the publishing platform that is operated by Sophia Publishing Group, founded in British Columbia of Canada. All Rights Reserved. Publisher BioSci Publisher
Computational Molecular Biology 2024, Vol.14, No.1 http://bioscipublisher.com/index.php/cmb © 2024 BioSci Publisher, registered at the publishing platform that is operated by Sophia Publishing Group, founded in British Columbia of Canada. All Rights Reserved. BioSci Publisher is an international Open Access publishing platform that publishes scientific journals in the field of bioscience registered at the publishing platform that is operated by Sophia Publishing Group (SPG), founded in British Columbia of Canada. Publisher BioSci Publisher Editedby Editorial Team of Computational Molecular Biology Email: edit@cmb.bioscipublisher.com Website: http://bioscipublisher.com/index.php/cmb Address: 11388 Stevenston Hwy, PO Box 96016, Richmond, V7A 5J5, British Columbia Canada Computational Molecular Biology (ISSN 1927-5587) is an open access, peer reviewed journal published online by BioSciPublisher. The Journal is publishing all the latest and outstanding research articles, letters, methods, and reviews in all areas of computational molecular biology, covering new discoveries in molecular biology, from genes to genomes, using statistical, mathematical, and computational methods as well as new development of computational methods and databases in molecular and genome biology. The papers published in the journal are expected to be of interests to computational scientists, biologists and teachers/students/researchers engaged in biology. All the articles published in Computational Molecular Biology are Open Access, and are distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. BioSciPublisher uses CrossCheck service to identify academic plagiarism through the world’s leading plagiarism prevention tool, iParadigms, and to protect the original authors’ copyrights.
Computational Molecular Biology (online), 2024, Vol. 14 ISSN 1927-6648 http://hortherbpublisher.com/index.php/cmb © 2024 BioSc iPublisher, registered at the publishing platform that is operated by Sophia Publishing Group, founded in British Columbia of Canada. All Rights Reserved. Latest Content 2024, Vol. 14, No.1 【Research Article】 Genomic Prediction and its Association with the Development of Dementia disease in the Elderly 1-8 Xiaojun Li, Shuiji Zhang DOI: 10.5376/cmb.2024.14.0001 Network Biology Reveals New Strategies for Understanding the Relationship Between Protein Function and Disease 28-35 Jiayao Zhou DOI: 10.5376/cmb.2024.14.0004 【Research Report】 The Application of Artificial Intelligence in Drug Discovery: Opportunities and Challenges 20-27 WeiWang DOI: 10.5376/cmb.2024.14.0003 【Review and Progress】 Artificial Intelligence and Drug Design: Future Prospects and Ethical Considerations 9-19 TaoChen DOI: 10.5376/cmb.2024.14.0002 Role of Proteomics in Unraveling Bacterial Virulence in Rice 36-44 Jianquan Li DOI: 10.5376/cmb.2024.14.0005
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 1 Review and Progress Open Access Genomic Prediction and its Association with the Development of Dementia disease in the Elderly Xiaojun Li, Shuiji Zhang Biotechnology Research Center, Cuixi Academy of Biotechnology, Zhuji, 311800, Zhejiang, China Corresponding author: jessi.j.zhang@foxmail.com Computational Molecular Biology, 2024, Vol.14, No.1 doi: 10.5376/cmb.2024.14.0001 Received: 29 Dec., 2023 Accepted: 30 Dec., 2023 Published: 04 Jan., 2024 Copyright © 2024 Li and Zhang, This is an open access article published under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Preferred citation for this article: Li X.J., and Zhang S.J., 2024, Genomic prediction and its association with the development of dementia disease in the elderly, Computational Molecular Biology, 14(1): 1-8 (doi: 10.5376/cmb.2024.14.0001) Abstract Dementia is a severe neurological disorder involving complex interactions between various genetic and environmental factors. This paper explores the association between genomic prediction and the development of dementia in the elderly. Through a systematic review of existing research, the study delves into genomics, the genetic basis of dementia, and the etiology related to the genome. The research further examines the methods and applications of genomic prediction, focusing on the use of polygenic risk scores and machine learning algorithms in dementia studies. Through case analyses of large-scale genomic studies, key genes associated with dementia, such as Alzheimer's disease, are revealed. Additionally, the paper thoroughly analyzes the major findings of existing research, emphasizing the filling of knowledge gaps and the provision of new insights. Finally, the paper discusses the challenges faced by genomic prediction, including methodological difficulties, challenges in data interpretation, ethical and privacy concerns, and more. Looking ahead to future research directions, the paper highlights the establishment of personalized genomic prediction models, the application of new technologies, and the potential value of genomic prediction in early diagnosis and prevention of dementia. Keywords Elderly dementia disease; Genomic prediction; Genetics; Polygenic risk scores; Machine learning algorithms Alzheimer's disease is a group of diseases mainly characterized by cognitive dysfunction, including Alzheimer's disease, vascular dementia, dancing disease, and frontotemporal dementia (Wu et al., 2021). According to statistics, Alzheimer's disease is an increasingly serious health problem among the elderly population worldwide, causing heavy burdens on patients and their families. With the trend of aging population, the incidence of Alzheimer's disease is on the rise, becoming an urgent problem to be solved in the medical field. The definition of Alzheimer's disease not only covers cognitive decline, but also includes the impact on individuals' daily living abilities. Epidemiological data of this condition show that its incidence is closely related to age and there are gender differences. According to the report of the World Health Organization, Alzheimer's disease has become a major health challenge for the elderly population worldwide (Fagundes et al., 2011). It is estimated that by 2050, the number of patients with Alzheimer's disease will exceed 200 million, posing a serious threat to the sustainability of the global health system (Nichols et al., 2022). Although in the past few decades, scientists have made significant progress in the etiology and pathophysiology of Alzheimer's disease (Simonetti et al., 2020), a radical cure has not yet been found. Therefore, more and more research is focused on understanding the genetic basis of Alzheimer's disease in order to intervene and treat it earlier. Due to the rapid development of genomics technology, genomic prediction has gradually become a popular direction in medical research. This method analyzes variations in an individual's genome to predict their risk of developing a specific disease. In the field of Alzheimer's disease, genomic prediction provides a new way to understand the role of genetic factors in the development of the disease (Oriol et al., 2019). Genomic prediction is a method based on genetic variation information to estimate individual susceptibility to a certain disease. Past studies have shown that Alzheimer's disease has a significant genetic predisposition, so using genomic prediction tools to explore its genetic basis has become the focus of scientists' attention.
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 2 With the continuous progress of high-throughput sequencing technology, researchers can more comprehensively interpret individual genomes and identify genetic variations associated with Alzheimer's disease. Genomic prediction has achieved some encouraging results in other disease areas, such as breast cancer and diabetes, providing reference and inspiration for the study of Alzheimer's disease. The purpose of this review is to systematically summarize the association studies between genomic prediction and the development of Alzheimer's disease, deeply analyze the main findings of existing literature, and explore future directions in this research field. This review will review the methods and applications of genomic prediction, focusing on the identified genetic variations associated with Alzheimer's disease and how these variations affect the development of the disease. Through this review, we aim to provide a comprehensive understanding of the genetics of Alzheimer's disease to the scientific community, and explore the potential applications of genomic prediction in early diagnosis, risk assessment, and personalized treatment. Ultimately, we hope to provide new theoretical support and research directions for the prevention and treatment of Alzheimer's disease. 1 Genomics and The Concept of Alzheimer's Disease 1.1 Definition of genomics Genomics, as a discipline that studies the entire genome, is an important branch in the field of biology (McGuire et al., 2020). The genome is the collection of all genes within an organism, and genes are DNA fragments that carry genetic information and are responsible for encoding proteins or regulating the expression of other genes. With the rapid development of technology, especially the application of high-throughput sequencing technology, it has become possible to fully understand the structure and function of the genome. The human genome consists of approximately 300 million base pairs and contains more than 20,000 genes. These genes carry instructions necessary for building and maintaining life, and their normal function is crucial for individual health. One of the main goals of genomics is to understand the arrangement, function, and interrelationships of genes in the genome. Whole-genome sequencing determines the location of genes, the proteins they encode, and regulatory elements that control gene expression. This provides researchers with an opportunity to delve deeper into the genome to reveal the molecular basis of various physiological and pathological processes in the body (Hu et al., 2021). In the study of Alzheimer's disease, the application of genomics can provide a more comprehensive understanding of patients' genetic information, especially those genetic variations associated with the development of Alzheimer's disease. By comparing the genomes of patients with those of healthy controls, scientists can identify specific genes, genomic regions, and variant types associated with Alzheimer's disease risk. These genomic findings provide powerful tools for Alzheimer's disease research, enabling researchers to delve deeper into the genetic basis, biological mechanisms, and potential therapeutic targets. 1.2 Genetic basis of Alzheimer's disease The genetic basis of Alzheimer's disease is complex and diverse, involving the interaction of multiple genes and environmental factors (Santos et al., 2020). The study of genetic factors helps to identify specific genes associated with Alzheimer's disease, providing important information for genomic prediction. Studies have found that some forms of Alzheimer's disease have a family history, indicating the role of genetics in the development of the disease. Although most cases are considered to be polygenic, there are also some forms of familial Alzheimer's disease that are associated with single-gene mutations, such as early-onset familial Alzheimer's disease. Alzheimer's disease, the most common form of dementia, has received extensive attention in terms of its genetic basis (Knopman et al., 2021). The genetic basis mainly involves genes related to amyloid precursor protein (APP),
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 3 steroid hormone receptors (APOE), and others. Mutations or variations in these genes may lead to abnormal aggregation of amyloid and neuronal damage, ultimately leading to cognitive decline. 1.3 Genome-related Alzheimer's disease gene In past studies, several genes have been identified as associated with the development of Alzheimer's disease. With technological advancements, large-scale genome-wide association studies (GWAS) have identified many new genes associated with Alzheimer's disease risk (Figure 1) (Bellenguez et al., 2022). Figure 1 Gene prioritization of related dementias (ADD) (Bellenguez et al., 2022) The APOE (apolipoprotein E) gene is one of the most closely associated genes with Alzheimer's disease. The ε4 allele of this gene is widely recognized as a major risk factor for Alzheimer's disease. Its mutant form is associated with increased amyloid deposition in the brain and neuronal damage, increasing the risk of developing Alzheimer's disease. The CLU (Clusterin) gene encodes the clusterin protein, which is involved in amyloid deposition and clearance in the brain. Multiple studies have found that variations in the CLU gene are associated with an increased risk of developing Alzheimer's disease. Clusterin is thought to play an important role in anti-inflammatory and neuroprotective functions in the brain, and its abnormalities can lead to inflammatory responses and neuronal death. The PICALM (Phosphatidylinositol Binding Clathrin Assembly Protein) gene encodes a protein that is involved in intracellular organelle transport in neurons. Variations in the PICALM gene are associated with an increased risk of developing Alzheimer's disease, possibly due to its role in promoting amyloid clearance and regulating neuronal survival (Ando et al., 2022).
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 4 Variations in the BIN1 (Bridging Integrator 1) gene are strongly associated with the risk of developing Alzheimer's disease (Gao et al., 2021). The BIN1 protein is involved in the regulation of membrane morphology, and its abnormalities can lead to amyloid deposition and neuronal damage. Some variations in the SORL1 (Sortilin-Related Receptor 1) gene have been found to be associated with an increased genetic risk of Alzheimer's disease. The SORL1 protein plays a role in amyloid clearance and trafficking, and its abnormalities can lead to abnormal amyloid aggregation. These genes are involved in multiple biological processes, and their mutations or variations can affect the pathogenesis of Alzheimer's disease through various pathways, including abnormal amyloid deposition, neuronal damage, and inflammatory responses. These findings provide a more comprehensive understanding of the pathogenesis of Alzheimer's disease and provide clues for exploring potential therapeutic targets. 1.4 The impact of the interaction between multiple genes and the environment The development of Alzheimer's disease is not only influenced by single genes, but also by the complex interactions between multiple genes and the environment. The introduction of polygenic risk scores (PRS) allows researchers to comprehensively consider the contribution of multiple genetic variants to disease risk. Additionally, environmental factors such as lifestyle, education level, psychosocial factors, etc. have also been found to be associated with the risk of Alzheimer's disease, and their interactions with genes further increase the complexity of the research. Researchers are working hard to reveal the interactions between these multiple factors to more comprehensively and accurately assess individual risk of Alzheimer's disease. Understanding these complex genetic and environmental interactions is crucial for developing prevention strategies and personalized treatment plans. 2 Methods and Applications of Genome Prediction 2.1 Basic principles of genome prediction Genome prediction is a method that predicts an individual's susceptibility to a certain disease by analyzing the genetic variations in their genome. The basic principle of this method is to establish a model that associates known genetic variations associated with the disease with disease risk, and then use this model to analyze the individual's genome data and estimate their likelihood of developing the disease. In genome prediction, commonly used methods include Polygenic Risk Scores (PRS) and machine learning algorithms (Lambert et al., 2019). PRS calculates a score by summing the risk weights of multiple genetic variants in an individual's genome, reflecting their overall genetic risk for a certain disease. Machine learning algorithms learn the genetic characteristics of the disease from a large amount of genome data, and then predict the risk for new individuals. 2.2 The application of genome prediction in Alzheimer's disease research As a cutting-edge technology, genome prediction has already shown great potential in Alzheimer's disease research. Its main areas of application include risk assessment, early diagnosis, and a deeper understanding of the genetics of Alzheimer's disease. By analyzing individual genome data, researchers can calculate the individual's likelihood of developing the disease, This enables more accurate personalized risk assessment. This is important for identifying high-risk populations, optimizing resource allocation, and developing personalized prevention strategies. For example, some researchers have successfully identified high-risk groups for Alzheimer's disease by constructing Polygenic Risk Scores (PRS) (Clark et al., 2022), providing strong support for personalized health management. By analyzing genome data, researchers can identify genetic markers associated with early lesions, providing the opportunity to identify patients before symptoms appear. This is crucial for early intervention, delaying disease progression, and improving treatment outcomes. Some studies have shown that combining genome prediction models with clinical symptoms can more accurately predict individuals' risk of developing the disease (Oriol et al., 2019), providing a new direction for early intervention and treatment.
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 5 Genome prediction also provides a new way to gain a deeper understanding of the mechanisms of Alzheimer's disease. In the process of studying the genes that play a key role in predictive models, researchers can explore the biological functions of these genes in the development and progression of the disease, providing clues for the discovery of new therapeutic targets. By comprehensively understanding the genetic variations associated with Alzheimer's disease, a better understanding of the pathogenesis of the disease can be achieved, providing support for the development of precision medicine. In summary, genome prediction has injected new vitality into Alzheimer's disease research and opened the door to the realization of individualized medicine. However, despite significant progress, a series of challenges still need to be faced, such as model complexity and data privacy issues. Future research needs to continuously improve genome prediction models, combining multi-source data to further improve prediction accuracy and reliability, and provide more powerful support for early intervention and treatment of Alzheimer's disease. 2.3 Case study analysis In the field of Alzheimer's disease, large-scale genomics studies provide opportunities for in-depth understanding of the genetic basis of the disease. Case study 1: Leonenko et al. (2019) developed a genome prediction model using GWAS data, focusing on Alzheimer's disease. They successfully integrated a large amount of genetic information, enabling more accurate prediction of individuals' risk of developing Alzheimer's disease. Notably, by combining genome prediction with clinical information, researchers not only improved prediction accuracy, but also provided new ways to distinguish high-risk individuals and develop early intervention plans. Case study 2: The meta-analysis study conducted by Jansen et al. (2019) has broadened our understanding of the genetic basis of Alzheimer's disease. They identified a series of new genetic variants associated with the risk of the disease, involving multiple functional pathways. This study not only provides new markers for the optimization of genome prediction models, but also deepens our understanding of the mechanism of Alzheimer's disease. Case study 3: Tan et al. (2017) introduced a new polygenic hazard score (PHS) method that is associated with amyloid and tau protein deposition in Alzheimer's disease. By focusing on these biological markers, researchers not only improved the precision of genome prediction, but also provided a new perspective for understanding the biological mechanism of Alzheimer's disease. The above studies, by integrating diverse genetic information and delving into biological markers, not only provide more comprehensive tools for individualized risk assessment, but also provide useful experiences for the development of future genome prediction models. These achievements lay a solid foundation for the prevention and treatment of Alzheimer's disease. 3 Genome Prediction and Its Association with Alzheimer's Disease 3.1 Comparison of research methods In the research on the association between genome prediction and Alzheimer's disease, different research teams have adopted different methods to reveal the relationship between genetic factors and the disease. There are certain similarities and differences among these research methods, which are mainly reflected in the following aspects: 1) Selection and weight assignment of genetic markers Different studies vary in the selection of genetic markers. Some studies focus on analyzing specific genes or gene regions, while others prefer to conduct comprehensive assessments through whole-genome approaches. Additionally, the weight assignment for different genes also varies among studies, which means that some studies may place more emphasis on the contribution of specific genes, while others may consider the role of multiple genes comprehensively.
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 6 2) Consideration of environmental factors The degree of consideration of environmental factors also varies among studies. Some studies incorporate environmental factors into the model, attempting to comprehensively consider the interaction between genetics and the environment, while others may primarily focus on genetic factors and treat them as the core of the research. 3) Selection of datasets and sample size The choice of datasets and sample size used in studies is also an important difference in research methods. Some studies may use data from different regions or ethnic groups to increase the external validity of the research, while others may focus on in-depth research within specific populations. 4) Analyzing the complexity of the model The complexity of the analytical models used in the study also varies. Some studies use relatively simple statistical models, while others may use more complex machine learning algorithms to better capture the potential patterns in the genomic data. Comparing the similarities and differences of research methods helps us to fully understand the diversity of genomic prediction research and provide guidance for future studies. By integrating different methods, we can build more comprehensive and robust genomic prediction models to better understand the genetic basis of Alzheimer's disease. 3.2 The application of the latest genome prediction technology in Alzheimer's disease research With the continuous progress of technology, the latest genome prediction technology has ushered in a new chapter in Alzheimer's disease research. One remarkable technology is single-cell RNA sequencing (scRNA-seq). Through this technology, researchers canin-depth exploration the differences in gene expression in brain tissue at the single-cell level, Revealed the unique contribution of different cell types in the development of Alzheimer's disease. This fine resolution allows us to more comprehensively and accurately understand the pathological process of Alzheimer's disease. In addition, the application of artificial intelligence (AI) has also brought revolutionary changes to genome prediction. Through the deep learning and pattern recognition capabilities of AI algorithms, researchers can discover hidden associations and patterns in large genomic datasets. In Alzheimer's disease research, AI can not only assist in analyzing the complex relationship between genes and diseases, but also predict the development trajectory of patient prognosis, providing higher-level guidance for precision medicine practice. The application of these emerging technologies allows researchers to explore the information of genomics at a deeper and more comprehensive level in Alzheimer's disease research. The introduction of these advanced technologies not only expands our understanding of the mechanism of the disease, but also points to the future development of genome prediction. In this promising field, it ispromising to break through the limitations of Alzheimer's disease research and provide more precise tools for the development of early diagnosis and treatment strategies. 3.3 Potential breakthroughs and discoveries By deeply exploring the genome data, researchers may identify specific genetic variations that emerge in the early stages of Alzheimer's disease development. These early predictive markers have the potential to become tools for earlier diagnosis of Alzheimer's disease, providing a window for early intervention and treatment. With a deeper understanding of the unique characteristics of individual genomes, doctors can develop more targeted treatment plans for patients. This means that in the future, it may be possible to develop more effective personalized medications and refined treatment plans that maximize therapeutic effectiveness and minimize the occurrence of side effects.
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 7 Additionally, with the widespread application of artificial intelligence in genomics, we are on the cusp of more intelligent genome prediction models. Through further optimization of machine learning algorithms, these models will be able to more accurately identify potential genetic associations and predict individual disease risks. This will provide clinicians with more reliable tools to better assist in decision-making and provide personalized medical advice. These findings will not only deepen our understanding of the mechanisms underlying Alzheimer's disease, but also hold promise for future treatment and prevention strategies. By combining the latest technologies and interdisciplinary research methods, significant achievements are expected in the near future. 4 Summary and Outlook Although genome prediction has made significant progress in Alzheimer's disease research, there are still a series of methodological challenges. The construction and optimization of models rely on large-scale genomic data. For complex diseases like Alzheimer's disease, larger and more diverse datasets are needed to improve the stability and generalization ability of models. Additionally, there are differences in methods and parameters used in different studies, which limits the consistency and comparability of results. Alzheimer's disease is a complex disease involving multiple factors and genetic markers, so the interpretability of genome prediction models becomes another challenge. Even if models perform well, we still have limited knowledge about the specific roles of specific genetic markers in the disease mechanism. This leads to difficulties in explaining prediction results, limiting the feasibility of genome prediction in clinical applications. Moreover, genome prediction models are typically based on population-level data, and there are significant individual differences. Personalized genome prediction models need to take into account individual lifestyles, environmental exposures, and other factors, which increases the complexity of data and interpretation. With the deepening of genomics research, ethical and privacy issues become increasingly important. Genome prediction involves a large amount of highly sensitive genetic information, so privacy protection needs to be strengthened during data collection, storage, and sharing. At the same time, how to explain and present the results of genome prediction to individuals, as well as how to apply this information in clinical practice, also requires clearer ethical guidance. It is also important to consider the social impact of genetic information. When conducting genome prediction, it may reveal information related to other diseases, traits, or family history, which may have potential impacts on individuals' employment, insurance, and other aspects. Establishing a sound ethical framework that safeguards individual rights while promoting scientific research is one of the important challenges facing current genome prediction research. Despite the challenges faced by genome prediction in Alzheimer's disease research, there are still many promising future research directions. Personalized genome prediction models will be the focus of future research. Researchers can also further explore the potential value of genome prediction in the early diagnosis and prevention of Alzheimer's disease. Additionally, with the continuous advancement of technology, the widespread application of new technologies such as whole-genome sequencing and single-cell sequencing will provide more abundant data for genome prediction research. The development of these technologies is expected to help researchers more comprehensively and deeply understand the genetic basis of Alzheimer's disease. Conflict of Interest Disclosure The authors affirm that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest. References Ando K., Nagaraj S., Küçükali F., De Fisenne M.A., Kosa A.C., Doeraene E., and Leroy K., 2022, PICALM and Alzheimer’s disease: an update and perspectives, Cells, 11(24): 3994. https://doi.org/10.3390/cells11243994
Computational Molecular Biology 2024, Vol.14, No.1, 1-8 http://bioscipublisher.com/index.php/cmb 8 Bellenguez C., Küçükali F., Jansen I.E., Kleineidam L., Moreno-Grau S., Amin N., and Goldhardt O., 2022, New insights into the genetic etiology of Alzheimer’s disease and related dementias, Nature genetics, 54(4): 412-436. Clark K., Leung Y.Y., Lee W.P., Voight B., and Wang L.S., 2022, Polygenic risk scores in Alzheimer’s disease genetics: methodology, applications, inclusion, and diversity, Journal of Alzheimer's Disease, 89(1):1-12. https://doi.org/10.3233/JAD-220025 Fagundes S.D., Silva M.T., Thees M.F.R.S., and Pereira M.G., 2011, Prevalence of dementia among elderly Brazilians: a systematic review, Sao Paulo Medical Journal, 129: 46-50. https://doi.org/10.1590/S1516-31802011000100009 Gao P., Ye L., Cheng H., and Li H., 2021, The mechanistic role of bridging integrator 1 (BIN1) in Alzheimer’s disease, Cellular and Molecular Neurobiology, 41(7): 1431-1440. https://doi.org/10.1007/s10571-020-00926-y Hu T., Chitnis N., Monos D., and Dinh A., 2021, Next-generation sequencing technologies: An overview, Human Immunology, 82(11): 801-811. https://doi.org/10.1016/j.humimm.2021.02.012 Jansen I.E., Savage J.E., Watanabe K., Bryois J., Williams D.M., Steinberg S., and Posthuma D., 2019, Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer’s disease risk, Nature genetics, 51(3): 404-413. Knopman D.S., Amieva H., Petersen R.C., Chételat G., Holtzman D.M., Hyman B.T., and Jones D.T., 2021, Alzheimer disease, Nature reviews Disease primers, 7(1): 33. https://doi.org/10.1038/s41572-021-00269-y Lambert S.A., Abraham G., and Inouye M., 2019, Towards clinical utility of polygenic risk scores, Human molecular genetics, 28(R2): R133-R142. https://doi.org/10.1093/hmg/ddz187 Leonenko G., Sims R., Shoai M., Frizzati A., Bossù P., Spalletta G., and Escott‐Price V., 2019, Polygenic risk and hazard scores for Alzheimer's disease prediction, Annals of clinical and translational neurology, 6(3): 456-465. https://doi.org/10.1002/acn3.716 McGuire A.L., Gabriel S., Tishkoff S.A., Wonkam A., Chakravarti A., Furlong E.E., and Kim J.S., 2020, The road ahead in genetics and genomics, Nature Reviews Genetics, 21(10): 581-596. https://doi.org/10.1038/s41576-020-0272-6 Nichols E., Steinmetz J.D., Vollset S.E., Fukutaki K., Chalek J., Abd-Allah F., .and Liu X., 2022, Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019, The Lancet Public Health, 7(2): e105-e125. https://doi.org/10.1002/alz.051496 Oriol J.D.V., Vallejo E.E., Estrada K., Peña J.G.T., and Alzheimer’s Disease Neuroimaging Initiative, 2019, Benchmarking machine learning models for late-onset alzheimer’s disease prediction from genomic data, BMC bioinformatics, 20. https://doi.org/10.1186/s12859-019-3158-x Santos C.D.S.D., Bessa T.A.D., and Xavier A.J., 2020, Factors associated with dementia in elderly, Ciencia & saude coletiva, 25: 603-611. https://doi.org/10.1590/1413-81232020252.02042018 Simonetti A., Pais C., Jones M., Cipriani M.C., Janiri D., Monti L., and Sani G., 2020, Neuropsychiatric symptoms in elderly with dementia during COVID-19 pandemic: definition, treatment, and future directions, Frontiers in psychiatry, 11: 579842. https://doi.org/10.3389/fpsyt.2020.579842 Tan C.H., Hyman B.T., Tan J.J., Hess C.P., Dillon W.P., Schellenberg G.D., and Desikan R.S., 2017, Polygenic hazard scores in preclinical Alzheimer disease, Annals of neurology, 82(3): 484-488. https://doi.org/10.1002/ana.25029 Wu J.W., Yaqub A., Ma Y., Koudstaal W., Hofman A., Ikram M.A., and Goudsmit J., 2021, Biological age in healthy elderly predicts aging-related diseases including dementia, Scientific reports, 11(1): 15929. https://doi.org/10.1038/s41598-021-95425-5
Computational Molecular Biology 2024, Vol.14, No.1, 9-19 http://bioscipublisher.com/index.php/cmb 9 Review and Progress Open Access Artificial Intelligence and Drug Design: Future Prospects and Ethical Considerations TaoChen Research Institute of Life Science, Jiyang College of Zhejiang A&F University, Zhuji, 311800, Zhejiang, China Corresponding email: 2693733238@qq.com Computational Molecular Biology, 2024, Vol.14, No.1 doi: 10.5376/cmb.2024.14.0002 Received: 4 Dec., 2023 Accepted: 7 Jan., 2024 Published: 18 Jan., 2024 Copyright ©2024 Chen, This is an open access article published under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Preferred citation for this article: Chen T., 2024, Artificial intelligence and drug design: future prospects and ethical considerations, Computational Molecular Biology, 14(1): 9-19 (doi: 10.5376/cmb.2024.14.0002) Abstract The rapid advancement of science and technology, artificial intelligence (AI) has penetrated into many fields and shown its great potential. In the field of drug design, the application of AI is gradually changing the traditional research and development model. This study first introduces the applicability of AI technology in drug design and its application examples at each stage, and analyzes its important role in improving R&D efficiency and success rate. Subsequently, the article looks forward to the future prospects of AI and drug design, including technological innovation, development trends, challenges and opportunities, and proposes corresponding development strategies. However, the widespread application of AI has also triggered many ethical considerations, such as data privacy, algorithm transparency, and definition of ethical responsibilities, which need to be treated with caution while promoting technological development. Finally, this study highlights how the relationship between innovation and ethics should be balanced in future research and makes corresponding recommendations. Keywords Artificial intelligence; Drug design; R&D efficiency; Future prospects; Ethical considerations In an era of rapid technological advancement, artificial intelligence (AI) has undoubtedly emerged as one of the most innovative and promising technological forces in the world. The research by Mak et al. (2019) profoundly revealed the disruptive application of AI in many fields such as image recognition, natural language processing, and data analysis. It not only greatly improved production efficiency, but also gave birth to profound changes in many industries. Today, the wave of AI has swept across like a sudden storm to drug design and discovery, a holy land involving professional fields such as gene sequence analysis, molecular docking, and drug effect prediction. With its excellent computing power, the wisdom of deep learning and the accuracy of pattern recognition, AI provides a new perspective and tools for drug design. It uses machine learning algorithms to efficiently analyze massive biological data, and uses natural language processing technology to parse medical literature and research results, providing unprecedented possibilities for drug discovery (Srivastav a et al., 2023). Looking back at the initial involvement of AI in the field of drug design, it mainly focused on auxiliary calculations and simulation experiments. However, with the rapid advancement of cutting-edge technologies such as deep learning and machine learning, AI has been deeply involved in every aspect of drug design, from target identification to molecular screening to the optimization of clinical trials. Its presence can be seen everywhere. This transformation not only significantly shortens the drug development cycle and reduces research and development costs, but more importantly, AI, with its unique insights and computing power, assists scientists in discovering drug candidates that are difficult to reach with traditional methods, and provides a basis for overcoming complex diseases. Provides new opportunities and hopes (Singh et al., 2023). However, just as all technological progress is accompanied by ethical challenges, the application of AI in the field of drug design also faces many ethical dilemmas. Issues such as data privacy, algorithm transparency, and responsibility attribution have gradually surfaced, triggering widespread concern and in-depth discussions from all walks of life. Therefore, this study aims to comprehensively examine the current application status of AI in the field of drug design, look forward to future development trends, and deeply explore the accompanying ethical considerations, with a view to providing valuable reference and guidance for scholars and practitioners in related fields.
Computational Molecular Biology 2024, Vol.14, No.1, 9-19 http://bioscipublisher.com/index.php/cmb 10 Through this study, we hope to further deepen our understanding of the role of AI in drug design and grasp its future development trends and potential challenges. At the same time, we also hope to trigger more in-depth thinking on how to balance technological innovation and ethical responsibility, and explore a path that not only promotes the healthy development of the field of drug design, but also respects ethics and morals. This not only has far-reaching significance in promoting progress in the field of drug design, but also provides useful reference and inspiration for the application of AI technology in wider fields. We firmly believe that under the dual guidance of technology and ethics, AI will create more miracles in the field of drug design and make greater contributions to human health. 1 Application of Artificial Intelligence Technology in Drug Design 1.1 Types of artificial intelligence technologies and their applicability in drug design Artificial intelligence technology covers multiple branches such as machine learning, deep learning, natural language processing, and reinforcement learning. Each technology has its unique application and applicability in drug design. For example, machine learning algorithms can be used to model the chemical structure and biological activity of known drug molecules, thereby predicting the potential activity of new molecules and guiding the design and optimization of drug molecules. In addition, machine learning can also be used to predict drug side effects, helping researchers avoid potential safety risks during the design stage. Wang et al. (2019) found that as drug design enters the era of big data, ML methods have gradually evolved into a deep learning (DL) method with stronger generalization capabilities and more effective big data processing, which further promotes the combination of artificial intelligence technology and computer-aided drug design technology promotes the discovery and design of new drugs. Zhong et al. (2018) found that deep learning technology can process more complex drug molecular structure information, such as three-dimensional conformation, intermolecular interactions, etc. By building a deep neural network model, deep learning can more accurately predict the interaction between drug molecules and targets, providing more accurate guidance for drug design. Thomas et al. (2022) discovered the antiviral drug Paxlovid designed for 3CL protease and the anti-tumor drug developed for KRAS protein. The success of these new drug discoveries all starts with the selection of targets and benefits from the assistance of AI technology. Natural language processing technology can assist scientific researchers in extracting useful information from massive documents and patents, such as the efficacy, side effects, and mechanisms of action of known drugs. Reinforcement learning is an artificial intelligence technology that learns interactively between an agent and the environment. By constructing a virtual environment that simulates the interaction between drug molecules and organisms, reinforcement learning algorithms can automatically explore and optimize the structure of drug molecules to maximize their effectiveness. efficacy and minimizing its side effects. Different artificial intelligence techniques have different applicability in drug design. Machine learning is suitable for modeling and predicting large amounts of data; deep learning is suitable for processing complex drug molecule structure information and making accurate predictions; reinforcement learning is suitable for optimizing the design process of drug molecules. 1.2 Application of artificial intelligence technology in various stages of drug design Artificial intelligence technology plays a vital role in all stages of drug design, bringing revolutionary changes to drug research and development. In the target identification stage, artificial intelligence technology helps researchers quickly and accurately identify potential drug targets related to specific diseases by analyzing big data such as genomics and proteomics. Hessle and Baringhaus (2018) found that deep neural networks showed improved predictability compared to baseline machine learning methods. At the same time, the scope of AI applications in early-stage drug discovery has expanded widely, such as de novo design of compounds and peptides and synthesis planning.
Computational Molecular Biology 2024, Vol.14, No.1, 9-19 http://bioscipublisher.com/index.php/cmb 11 In the molecular screening and optimization stage, artificial intelligence technology helps researchers quickly select potentially active candidate molecules from a huge compound library by building prediction models, and optimizes the structure of these molecules to improve their efficacy and reduce side effects. For example, the virtual screening method based on machine learning (Figure 1) can use the chemical structure and biological activity data of known active molecules to build a prediction model to predict and rank the activity of new molecules, thereby quickly screening out candidate molecules with potential activity. Figure 1 Virtual screening (Zhang et al., 2024) Artificial intelligence technology also plays an important role in the clinical trial stage. For example, DeepMind collaborated with Moorfields Eye Hospital to develop the Streams system, which uses deep learning technology to analyze eye scan images, automatically identify and interpret complex images, and provide preliminary diagnostic recommendations (https://zhuanlan.zhihu.com/p/41970785). This technology helps solve the problem of scarce expert resources and allows patients to receive timely diagnosis. In addition, prediction models based on machine learning can also predict and evaluate the results of clinical trials, providing strong support for the design and optimization of clinical trials. These applications can not only improve the efficiency and accuracy of clinical trials, but also help reduce the costs and risks of clinical trials. Zhang et al. (2022) found that despite a large investment of money and time, the success rate of clinical testing is still less than 15%. Approximately 50% of drug discovery failures are due to poor pharmacokinetic properties (absorption, distribution, metabolism, excretion and toxicity). With the development of computational methods, the speed and success rate of drug discovery have greatly improved.
Computational Molecular Biology 2024, Vol.14, No.1, 9-19 http://bioscipublisher.com/index.php/cmb 12 1.3 Analysis of the impact of artificial intelligence technology on drug design efficiency and success rate Artificial intelligence technology has had a profound impact on the efficiency and success rate of drug design, and has greatly promoted progress in the field of drug research and development. At all stages of drug design, artificial intelligence technology has significantly improved work efficiency. The traditional drug design process requires a lot of manual experiments and data analysis, which is time-consuming and labor-intensive. Artificial intelligence technology can quickly process and analyze large-scale data sets through automated and intelligent methods, thus greatly shortening the time cycle of drug design (Moingeon et al., 2022). For example, Exscientia cooperates with Japan's Sumitomo Dainippon Pharma to use artificial intelligence platforms to automatically generate and screen drug molecules, accelerating the drug discovery process. Public data shows that this technology shortens drug development time from 5-10 years to 1-2 years, improving the success rate. During the cooperation, a number of innovative drug candidates in cancer, neurological diseases and other fields have been discovered and entered the clinical trial stage (https://zhuanlan.zhihu.com/p/114953741). Artificial intelligence technology has also significantly improved the success rate of drug design. In the traditional drug design process, there is often a high failure rate due to limitations in experimental conditions, data quality, analysis methods and other factors. Artificial intelligence technology can screen potential drug candidates at an early stage through accurate data analysis and prediction models, thereby reducing the risk of later experimental failure. Zhang Minquan et al. (2024) found that artificial intelligence technology uses big data to screen out corresponding compounds for molecular simulation, and feeds the simulation results back to the artificial intelligence system for learning, and continuously optimizes the artificial neural network. The combined use of artificial intelligence and molecular simulation technology improves the efficiency of drug design research, reduces the impact of human factors on simulation results, and increases the credibility of simulation results. For example, in the preclinical research stage, artificial intelligence can use machine learning algorithms to accurately predict the biological activity, pharmacokinetic properties, and toxicity of candidate drugs, helping researchers discover potential problems in advance and optimize them, thus improving the quality of drugs. Success rate in entering clinical trials. In addition, artificial intelligence can also discover biomarkers and risk factors closely related to patient efficacy and safety by mining and analyzing clinical trial data, providing strong support for the design and optimization of clinical trials, and further improving the success rate of drug development. 2 Ethical Considerations in Artificial Intelligence and Drug Design 2.1 Data privacy and security issues In the process of applying artificial intelligence to drug design, data privacy and security issues are particularly critical. This involves how to reasonably and legally collect, store and use large amounts of biometric data, medical information, patient records and other sensitive content. The protection of data privacy is a dual ethical and legal requirement. Murdoch (2021) research stated that patients' personal information, genetic data, etc. are highly sensitive information, and once leaked, it may have a serious impact on the patient's life, work, and even personal safety. Therefore, when collecting these data, the patient’s explicit consent must be obtained and their rights to information, choice, and refusal must be fully respected. At the same time, the data storage and transmission process also requires strict encryption to prevent data from being illegally obtained or abused. Pesapane et al. (2018) analyzed the legal framework regulating medical devices and data protection in Europe and the United States, assessed the developments currently taking place, and stated that data security issues cannot be ignored. During the drug design process, large amounts of data need to be shared and exchanged between different institutions, platforms and even countries. This brings great challenges to data security. On the one hand, it is necessary to establish a complete data sharing mechanism to ensure that data flows under the premise of legality
Computational Molecular Biology 2024, Vol.14, No.1, 9-19 http://bioscipublisher.com/index.php/cmb 13 and compliance; on the other hand, it is also necessary to strengthen the supervision of the data sharing process to prevent data from being tampered with, abused or used for other illegal purposes (Zhang et al., 2022). With the continuous development of artificial intelligence technology, the value and importance of data have become increasingly prominent. This makes data a target for various attacks and theft. Therefore, it is necessary to continuously improve data security protection capabilities and adopt the latest technical means and methods to deal with various network attacks and data leakage incidents. 2.2 Transparency and explainability of artificial intelligence decision-making In the field of drug design, the transparency and explainability of artificial intelligence decision-making have become the focus of public attention, scientific researchers and regulatory agencies. This is not only because of the advancement of technology, but also because AI decision-making is directly related to human health and life safety. On August 15, 2021, Professor Liang Zheng, Vice Dean of the Institute of International Governance of Artificial Intelligence at Tsinghua University, attended "The 4th Issue of the Future Forum AI Ethics and Governance Series - Reliability and Explainability of AI Decision-Making". Professor Liang Zheng pointed out that reliable AI should have four major elements: security, fairness, transparency, and privacy protection. Therefore, "trustworthiness" and "explainability" are positively related, especially for users and the public. Implementing algorithm explainability is an important part of ensuring reliability and trust (https://aiig.tsinghua.edu.cn/ info/1296/1328.htm). Transparency, simply put, refers to the extent to which the processes and logic behind AI decisions can be understood and viewed. In drug design, an AI model might recommend a certain molecular structure as a potential drug candidate based on millions of data points and complex algorithms. But the question is, how are these decisions made? What data is it based on? What algorithms are used? Deng et al. (2022) studied common data resources, molecular representations, and benchmark platforms to decompose artificial intelligence technology into model architectures and learning paradigms. Reflects the technical development of artificial intelligence in drug discovery over the years and provides a GitHub repository containing a series of papers (and code, if applicable) as a learning resource, which is updated regularly. These need to be made clear. Transparency requires that the AI system can provide sufficient information to allow external observers to understand the basis and logic of its decisions. Explainability goes a step further, requiring AI to not only demonstrate its decision-making process, but also explain the reasons for its decisions in a way that humans can understand. In drug design, this means that AI needs to be able to explain why a certain molecular structure was chosen and not others. This explanation cannot be just "because the algorithm says so", but should be based on specific chemical or biological principles or known experimental results. However, achieving transparency and explainability is not easy. The decision-making process of AI often involves large amounts of data and complex calculations, which is difficult to describe in simple language. In addition, some AI models themselves are "black box" models, and their internal logic is not easy to understand. Therefore, researchers need to continuously explore new methods and technologies to improve the transparency and explainability of AI decision-making (Schneider et al., 2020). 2.3 Definition of ethical responsibilities of artificial intelligence in drug design In the field of drug design, the application of artificial intelligence is becoming more and more widespread. However, it is followed by a series of complex ethical issues, the most core of which is the definition of ethical responsibility. The ethical responsibility of artificial intelligence in drug design is not a simple "yes or no", but a complex issue involving multiple levels and requiring careful consideration. It must be recognized that although artificial intelligence has powerful computing power and data analysis capabilities, it is still a tool based on human programming and algorithms. Therefore, Jing et al. (2018) found that
RkJQdWJsaXNoZXIy MjQ4ODYzNA==