In recent years, the application of artificial intelligence has been continuously expanding across various fields, with molecular medicine being a particularly noteworthy and researched domain. Molecular medicine, as a critical branch of medicine, focuses on studying diseases at the molecular level, including diagnosis, treatment, and prevention, providing new methods and strategies for human health. Meanwhile, artificial intelligence, with its powerful algorithms and models, is gradually revealing significant potential in the field of molecular medicine.

Understanding the occurrence and development of diseases at the molecular level is the foundation for comprehending the essence of diseases and finding treatment strategies. However, traditional approaches in the past often faced limitations in data scale and complexity. The emergence of artificial intelligence technology has provided us with a new avenue to overcome these traditional constraints. By utilizing artificial intelligence algorithms such as machine learning, deep learning, and neural networks, we can process and analyze large-scale biological data, extract valuable information, and identify key molecular markers associated with diseases. This capability offers researchers and healthcare professionals a more comprehensive and in-depth understanding of diseases.

Understanding the occurrence and development of diseases at the molecular level is the foundation for comprehending the essence of diseases and finding treatment strategies. However, traditional approaches in the past often faced limitations in data scale and complexity. The emergence of artificial intelligence technology has provided us with a new avenue to overcome these traditional constraints. By utilizing artificial intelligence algorithms such as machine learning, deep learning, and neural networks, we can process and analyze large-scale biological data, extract valuable information, and identify key molecular markers associated with diseases. This capability offers researchers and healthcare professionals a more comprehensive and in-depth understanding of diseases.

However, despite the tremendous potential demonstrated by artificial intelligence in molecular medicine, it also faces a series of challenges and issues. Among them, issues of data quality, the need for data privacy protection, as well as the interpretability and reliability of algorithms are urgent problems that need to be addressed. At the same time, the application of artificial intelligence in clinical settings also needs to overcome obstacles such as recognition and acceptance.

This review aims to comprehensively discuss the application prospects of artificial intelligence in molecular medicine. The review will introduce the basic principles and methods of artificial intelligence in molecular medicine, as well as its potential applications in disease diagnosis, drug development, and treatment. Simultaneously, discussions will also be conducted on the current challenges and limitations, and corresponding solutions will be proposed. We believe that by fully harnessing the advantages of artificial intelligence technology, we can propel the development of molecular medicine, bringing new breakthroughs and advancements in the diagnosis, treatment, and prevention of diseases.

1 Basic Principles of Artificial Intelligence in Molecular Medicine

1.1 Concept and classification of artificial intelligence

Artificial intelligence (AI) refers to a discipline that utilizes technologies such as computer science and machine learning to simulate and mimic human intelligence. In the field of molecular medicine, AI is widely used in areas such as data analysis, pattern recognition, and decision support (Liu and Chen, 2021). Based on different features and tasks, AI can be classified into expert systems, machine learning, deep learning, natural language processing, and reinforcement learning.

Expert systems are knowledge- and rule-based approaches that simulate the decision-making process of experts by encoding their experience and knowledge into computer programs. In molecular medicine, expert systems can be employed to assist in tasks such as disease diagnosis, drug design, and treatment plan selection.

Machine learning is a data-driven method that enables computers to make predictions and decisions by learning patterns and patterns from a large amount of data. It is a sub direction of  artificial intelligence. In molecular medicine, machine learning can be used for tasks such as identifying molecular markers, predicting disease risks, and optimizing drug combinations (Anuraj et al., 2021). Additionally, deep learning, a branch of machine learning (Figure 1), involves constructing and training models with multiple layers of neural networks to achieve advanced abstraction and processing of complex data. In molecular medicine, deep learning can be applied to tasks such as image analysis, interpretation of gene expression data, and prediction of protein structures.

Figure 1 Relationship between artificial intelligence and related concepts

Natural language processing is a technology that converts human language into computer understandable and processable language. In molecular medicine, natural language processing can be used to analyze and understand textual information such as literature, medical records, and clinical guidelines, thereby extracting and inferring medical knowledge. Reinforcement learning is a method of learning through trial and error and reward and punishment mechanisms. In molecular medicine, reinforcement learning can be used to optimize personalized treatment plans, develop intelligent drug delivery systems, and simulate molecular interactions.

1.2 Basic concepts and research contents of molecular medicine

Molecular medicine is an interdisciplinary field based on research in fields such as molecular biology, genomics, and medicine, aiming to gain a deeper understanding of the molecular mechanisms of diseases and utilize them for disease diagnosis, prevention, and treatment. Researchers in molecular medicine focus on the function and regulation of genes and genomes, as well as the association between gene mutations and disease occurrence. In addition, finding and researching molecular markers related to diseases is also an important aspect, which can be used for disease diagnosis and prognosis prediction.

Molecular medicine also utilizes molecular diagnostic methods such as gene sequencing and protein testing for early and accurate diagnosis of diseases (Hu and Wei, 2021). At the same time, molecular medicine is also committed to drug development and personalized treatment, by studying the interaction between drugs and target molecules, developing new drugs, and implementing personalized treatment strategies based on individual genomic information. Molecular medicine combines molecular biology and imaging techniques to visualize and evaluate the activity and distribution of disease occurrence areas through the use of labeled molecules. Overall, the goal of molecular medicine is to utilize molecular level knowledge and technology to gain a deeper understanding of diseases, thereby providing scientific basis for personalized medicine.

2 The History and Development of Gene Therapy

2.1 Early disease diagnosis

Artificial intelligence (AI) has a wide range of applications in molecular medicine research, especially playing an important role in the early diagnosis of diseases. It can process a large amount of molecular biology, genomics, and clinical data through data analysis and pattern recognition, helping molecular medical researchers discover the correlation between diseases and molecular features. Additionally, AI can also be applied to the processing and analysis of biological images, using deep learning algorithms to swiftly analyze biological tissue slices or cell images. This facilitates the early detection of cancer by providing opportunities for identifying tumor targets (Figure 2) and supporting physicians in early cancer diagnosis (Bhinder et al., 2021).

Figure 2 Application of artificial intelligence in tumor detection, diagnosis, and treatment

Furthermore, artificial intelligence can discover and validate molecular biomarkers by integrating diverse data to identify disease-related markers and conduct rapid validation. AI can also utilize individual genomic information, clinical data, and environmental factors to establish models for assessing an individual's risk of developing a particular disease, providing personalized disease prevention and intervention strategies for individuals. In summary, AI is widely applied in molecular medicine research, particularly in early disease diagnosis, offering researchers new methods and tools.

2.2 Recognition and prediction of molecular markers

Molecular markers refer to molecular features associated with specific diseases or physiological states, such as genes, proteins, or metabolic products. The application of artificial intelligence (AI) technologies enables the rapid and accurate identification and prediction of these molecular markers (Ullah et al., 2020). The application of AI in molecular medicine research plays an important role in the recognition and prediction of molecular markers. By processing and analyzing large-scale molecular biology data and conducting integrated analyses of different data sources, AI assists researchers in gaining a better understanding of the molecular mechanisms underlying diseases and provides a scientific basis for early diagnosis and personalized treatment.

Artificial intelligence can utilize machine learning and deep learning algorithms to analyze large-scale molecular biology data, such as genomic, transcriptomic, and proteomic data. By processing and pattern recognition of these data, artificial intelligence can uncover molecular features associated with diseases. For instance, through the analysis of extensive genomic data, artificial intelligence can identify genetic mutations associated with the occurrence and development of diseases. This aids in determining potential therapeutic targets and predicting the efficacy of drugs for specific individuals.

Furthermore, artificial intelligence can leverage molecular image processing and analysis techniques to diagnose and classify diseases by analyzing the expression patterns of molecular markers in biological tissue slices or cell images. The comprehensive analysis of multiple data sources is also a powerful application of artificial intelligence in the identification and prediction of molecular markers. By integrating data from various dimensions, such as genomic, transcriptomic, proteomic data, as well as clinical manifestations, artificial intelligence can discover more precise and effective molecular markers, aiding in predicting the occurrence and development of diseases.

2.3 Drug design and discovery

The application of artificial intelligence in molecular medicine research is extensive, with a significant role in drug design and discovery. Artificial intelligence can utilize machine learning and deep learning algorithms to analyze vast datasets of compounds and drug activity, thereby enabling drug screening and virtual screening to rapidly assess the potential activity and properties of candidate compounds. Additionally, through algorithms like generative adversarial networks, artificial intelligence can generate new drug molecules, expand chemical space and search for new drugs with better activity and selectivity.

Artificial intelligence can predict the side effects and toxic effects of drugs, aiding in optimizing drug design and developing safer and more effective medications. By predicting and simulating the structure and function of proteins, artificial intelligence can also enhance the activity and selectivity of drugs. Furthermore, utilizing machine learning and data mining techniques, artificial intelligence can identify synergistic drug combinations to improve therapeutic outcomes. In summary, the application of artificial intelligence in molecular medicine research provides new perspectives and tools for drug design and discovery, accelerating the drug development process and offering greater possibilities for disease treatment (Liu, 2018).

2.4 Treatment response and personalized therapy

The application of artificial intelligence in molecular medicine research extends to the realms of treatment response and personalized therapy. By analyzing multidimensional information such as clinical data, biomarker data, and genomic data, artificial intelligence can establish predictive models to forecast patients' responses to specific treatment modalities, thereby enabling personalized therapy. Additionally, through the integration of extensive clinical trial data, medical literature, and actual clinical data, artificial intelligence conducts knowledge discovery and pattern recognition. This provides physicians with guidance on the effectiveness and side effects of different treatment approaches, assisting in the optimization of treatment plans.

Artificial intelligence can also utilize patients' genomic data to predict their responses to different drugs, offering personalized drug selection and dosages. By continuously monitoring patients' health data in real-time and providing personalized treatment recommendations, artificial intelligence can assist patients and healthcare professionals in promptly adjusting treatment plans in a timely manner for more precise personalized therapy (Boniolo et al., 2021). The application of artificial intelligence in molecular medicine research provides new perspectives and methods for treatment response and personalized therapy, fostering the implementation of personalized medicine in the healthcare field and providing better medical outcomes for patients.

3 Opportunities of Artificial Intelligence in Molecular Medicine

3.1 Enhance research efficiency and accuracy

Artificial intelligence presents significant opportunities in molecular medicine research aimed at improving research efficiency and accuracy. Through data analysis and mining, AI can efficiently process large-scale molecular biology data, spanning from genomics to proteomics and metabolomics data, thereby accelerating the research process. AI plays a crucial role in drug screening and virtual screening, utilizing machine learning algorithms to rapidly assess the activity and properties of candidate compounds, providing accurate references for drug research.

Additionally, the application of artificial intelligence enables the prediction and simulation of protein structures and functions, facilitating the design and development of novel drugs. Through pattern recognition and disease classification, AI contributes to improving the accuracy of early disease diagnosis and treatment selection, realizing personalized therapy. Real-time monitoring and prediction constitute another critical area where AI technology can continuously monitor patients' physiological parameters and health data, establishing predictive models to forecast disease progression and treatment outcomes for timely adjustments to treatment plans. The application of artificial intelligence in molecular medicine research provides extensive opportunities to improve research efficiency and accuracy, promising substantial advancements in medical research.

3.2 Accelerating drug research and development

Artificial intelligence presents numerous opportunities in molecular medicine, one of which is to accelerate the process of the drug research and development (Huang et al., 2021). Through efficient data mining and machine learning algorithms, AI can rapidly screen and evaluate the activity and properties of a large number of candidate compounds, thereby improving the speed and success rate of drug development. Additionally, AI can predict and simulate the interaction between drugs and targets, guiding drug design and optimization to discover more active and selective drug molecules. Moreover, data-driven drug repositioning methods can apply existing drugs to new indications, saving time and costs.

Personalized drug therapy represents another significant opportunity, where artificial intelligence utilizes patients' genomic data and clinical information to provide personalized treatment recommendations for physicians, aiming to enhance efficacy and reduce side effects. Furthermore, AI can predict drug interactions and side effects, assisting physicians in assessing the safety and selectivity of medications. In summary, artificial intelligence has accelerated the process of drug development and transformation in molecular medicine, providing robust support for the smooth development of new drugs and personalized treatment.

3.3 Advancing precision medicine

Artificial intelligence offers numerous opportunities in advancing precision medicine within molecular medicine. Genomics stands out as a crucial domain among them. By analyzing patients' genomic data, artificial intelligence can identify gene variations and mutations associated with diseases, providing support for early detection and personalized treatment. Additionally, AI can predict drug responsiveness, assisting physicians in selecting the most suitable drug treatment plans. Another significant opportunity lies in data analysis. Through the efficient data mining and analysis algorithms, artificial intelligence can uncover patterns and trends in clinical data, physiological parameters, and genomic data, aiding physicians in precise diagnosis and treatment decision-making. In the realm of imaging diagnostics, AI's deep learning algorithms can automatically identify and analyze lesions and abnormalities in medical images, assisting in early screening and accurate diagnosis.

The application of artificial intelligence in follow-up and monitoring also holds immense potential. By real-time monitoring of patients' physiological parameters and health data, coupled with AI analysis and prediction, personalized follow-up and monitoring services can be provided to optimize treatment plans and prevent disease deterioration. Furthermore, AI plays a crucial role in health management and prediction. Through data analysis, AI can establish health prediction models, proactively forecasting potential health risks and disease development trends for patients, thus supporting personalized prevention and management. In conclusion, artificial intelligence in molecular medicine provides numerous opportunities for achieving precision medicine. Through applications in genomics, data analysis, imaging diagnostics, follow-up and monitoring, as well as health management and prediction, personalized healthcare services can be achieved, enhancing the precision of diagnosis and treatment, and delivering more effective health management and preventive measures for patients (Filipp et al., 2019).

4 Challenges of Artificial Intelligence in Molecular Medicine

4.1 Data quality and quantity

While the application of artificial intelligence in molecular medicine presents numerous opportunities for achieving precision medicine, it also encounters several challenges. Among these, data quality and quantity are two crucial aspects.

Data quality poses a challenge for artificial intelligence in molecular medicine. The collection and processing of medical data involve multiple sources and institutions, introducing the possibility of errors, omissions, or biases. Ensuring the accuracy, consistency, and completeness of data is paramount to prevent inaccurate results in the analysis and prediction by models. Additionally, privacy and security issues need thorough consideration to protect the privacy and security of medical data.

Data quantity represents another critical challenge. Acquiring high-quality data in the field of molecular medicine is particularly difficult, especially for certain diseases or rare conditions where data may be more limited. The lack of sufficiently diverse and representative dataset constrains the performance and applicability of artificial intelligence models. Therefore, actively promoting data sharing and collaboration has become crucial.

To overcome these challenges, it is necessary to strengthen data quality management and standardization to ensure the accuracy and completeness of data, along with measures to protect data privacy and security. Additionally, actively promoting data sharing and collaboration can increase the quantity and diversity of data. Establishing mechanisms for cross-institutional and interdisciplinary data collaboration can facilitate data sharing and integration, enhancing the performance and applicability of artificial intelligence models.

In conclusion, the development of artificial intelligence in molecular medicine faces challenges in terms of data quality and quantity. By strengthening data quality management, privacy protection, and data sharing, these challenges can be overcome, fostering the application and advancement of artificial intelligence in the field of molecular medicine, and making greater contributions to the realization of precision medicine.

4.2 Privacy and ethical issues

The application of artificial intelligence in molecular medicine has brought significant opportunities, yet it also faces challenges in terms of privacy and ethics. Individual privacy emerges as a critical concern, as medical data encompasses sensitive personal information, posing risks of privacy breaches if not handled appropriately. Therefore, ensuring the security and confidentiality of medical data is paramount, necessitating reasonable measures in data standardization, encryption, and storage (Zhang and Xu, 2020).

Furthermore, the use of artificial intelligence in molecular medicine raises ethical and legal considerations regarding data usage. When analyzing and predicting with medical data, it is imperative to ensure the legality and compliance of data usage, adhering to applicable laws and regulations. Additionally, ethical considerations should be taken into account to prevent the misuse of data or any inappropriate impacts. For data related to specific groups or individuals, principles of informed consent and anonymization should be followed to respect individual privacy rights.

Addressing privacy and ethical concerns requires a series of measures. Establishing rigorous privacy protection measures and security management mechanisms ensures that medical data is not abused or leaked. Simultaneously, creating transparent data usage policies and procedures empowers individuals with the right to choose and control their own data. Legitimate and compliant data access controls and quality management mechanisms guarantee the reliability and credibility of the data.

In conclusion, privacy and ethical issues related to artificial intelligence in molecular medicine demand attention. By implementing stringent privacy protection measures, adhering to laws and regulations, enhancing transparency, and ensuring the compliance of data usage, we can address these issues and protect the privacy and ethical integrity of the data. Only by addressing privacy and ethical concerns responsibly can artificial intelligence contribute effectively to innovation and progress in the field of molecular medicine.

5 Summary and Outlook

The application of artificial intelligence in the field of molecular medicine has brought enormous opportunities, but it is also accompanied by a series of challenges. Artificial intelligence can accelerate innovation in fields such as genomics, protein research, and new drug discovery. Through deep learning and data analysis techniques, potential new treatment methods and drug targets can be quickly identified. Artificial intelligence plays an important role in analyzing patient genomic data, providing personalized medical decisions and treatment plans.

However, artificial intelligence in molecular medicine also faces several challenges. Data privacy and security are crucial concerns, as medical data involves individual privacy that needs to ensure the security and confidentiality of the data to prevent unauthorized access and abuse.. The interpretability of the model is a key issue. Despite the high accuracy of artificial intelligence models in prediction and analysis, the lack of interpretability limits their application in clinical practice. This makes it challenging for doctors and patients to understand the basis and mechanisms behind the predictive results.

With the continuous development of technology, we can expect further breakthroughs in the application of artificial intelligence in the field of molecular medicine. To address these challenges, there is a need to strengthen privacy protection and data security measures, establish stringent data privacy policies and security standards, and employ encryption and storage technologies to ensure the security of medical data while safeguarding individual privacy in data sharing.

In terms of enhancing model interpretability, the development of explainable artificial intelligence technologies can facilitate a better understanding of the decision-making processes and predictive results of models. This assists doctors and patients in evaluating and trusting the predictive results, increasing their acceptability in clinical applications. In promoting interdisciplinary collaboration, strengthening collaboration across the fields of medicine, biotechnology, and artificial intelligence is essential to collectively address the challenges in the field of molecular medicine. By integrating professional knowledge and technology from different fields, innovation and development can be driven.

In summary, the potential of artificial intelligence in molecular medicine is enormous, but it also faces challenges related to privacy and model interpretability. By strengthening privacy protection, improving model interpretability, and fostering interdisciplinary collaboration, artificial intelligence technology can be better utilized to promote the development of molecular medicine and providing more opportunities for medical innovation and patient health.

Acknowledgments

I would like to express my gratitude to Ms. Fang Keyan for her guidance and review of this paper. Her constructive feedback and numerous valuable suggestions have greatly contributed to the refinement of this research.

References

Anuraj N., Ravina K., Poonam T., Maddala M., Diksha S., Garima T., Alejandro S.P., and Sanjeev K.S., 2021 Artificial intelligence, big data and machine learning approaches in precision medicine & drug discovery, Current Drug Targets, 22(6): 631-655.

https://doi.org/10.2174/18735592MTEzsMDMnz

PMid:33397265

Boniolo F., Dorigatti E., Ohnmacht A.J., Saur D., Schubert B., and Menden M.P., 2021, Artificial intelligence in early drug discovery enabling precision medicine, Expert Opinion on Drug Discovery, 16(9): 991-1007.

https://doi.org/10.1080/17460441.2021.1918096

PMid:34075855

Bhinder B., Gilvary C., Madhukar N.S., and Elemento O., 2021, Artificial intelligence in cancer research and precision medicine, Cancer Discov, 11 (4): 900-915.

https://doi.org/10.1158/2159-8290.CD-21-0090

PMid:33811123 PMCid:PMC8034385

Filipp F.V., 2019, Opportunities for artificial intelligence in advancing precision medicine, Curr. Genet Med. Rep., (7): 208-213.

https://doi.org/10.1007/s40142-019-00177-4

PMid:31871830 PMCid:PMC6927552

Huang F., Yang H.F., and Zhu X., 2021, Progress in the application of artificial intelligence in new drug discovery, Yaoxue Jinzhan (Progress in Pharmaceutical Sciences), 45(7): 502-511.

Hu H., and Wei W., 2021, Methods of molecular diagnosis and significance in breast cancer, Linchuang Waike Zazhi (Journal of Clinical Surgery), 29(3): 287-290.

Liu Q., 2018, Artificial intelligence and drug discovery, Dier Junyi Daxue Xuebao (Academic Journal of Second Military Medical University), 39(8): 869-872.

Liu Z.H., and Chen T.Y., 2021, Artificial intelligence: a new driving force for the development of medicine, Shengzhangbing Yu Touxi Shengyizhi Zazhi (Chinese Journal of Nephrology, Dialysis & Transplantation), 30(1): 1-6.

Ullah M., Akbar A., and Yannarelli G., 2020, Applications of artificial intelligence in, early detection of cancer, clinical diagnosis and personalized medicine, Artif Intell Cancer, 1(2): 39-44.

https://doi.org/10.35713/aic.v1.i2.39

Zhang R., Xu F., 2020, Research on the ethical issues of artificial intelligence medicine and its countermeasures, Yixue Yu Zhexue (Medicine and Philosophy), (13): 14-19.