Pokushalov Evgeny, Ponomarenko Andrey, Shrainer Evgenya, Kudlay Dmitry, Miller Richard
Center for New Medical Technologies, Novosibirsk 630090, Russia.
Scientific Research Laboratory, Triangel Scientific, San Francisco, CA 94101, USA.
Nutrients. 2024 Nov 25;16(23):4033. doi: 10.3390/nu16234033.
Dietary supplements (DS) are widely used to address nutritional deficiencies and promote health, yet their indiscriminate use often leads to reduced efficacy, adverse effects, and safety concerns. Biomarker-driven approaches have emerged as a promising strategy to optimize DS prescriptions, ensuring precision and reducing risks associated with generic recommendations. This narrative review synthesizes findings from key studies on biomarker-guided dietary supplementation and the integration of artificial intelligence (AI) in biomarker analysis. Key biomarker categories-genomic, proteomic, metabolomic, lipidomic, microbiome, and immunological-were reviewed, alongside AI applications for interpreting these biomarkers and tailoring supplement prescriptions. Biomarkers enable the identification of deficiencies, metabolic imbalances, and disease predispositions, supporting targeted and safe DS use. For example, genomic markers like MTHFR polymorphisms inform folate supplementation needs, while metabolomic markers such as glucose and insulin levels guide interventions in metabolic disorders. AI-driven tools streamline biomarker interpretation, optimize supplement selection, and enhance therapeutic outcomes by accounting for complex biomarker interactions and individual needs. Despite these advancements, AI tools face significant challenges, including reliance on incomplete training datasets and a limited number of clinically validated algorithms. Additionally, most current research focuses on clinical populations, limiting generalizability to healthier populations. Long-term studies remain scarce, raising questions about the sustained efficacy and safety of biomarker-guided supplementation. Regulatory ambiguity further complicates the classification of supplements, especially when combinations exhibit pharmaceutical-like effects. Biomarker-guided DS prescription, augmented by AI, represents a cornerstone of personalized nutrition. While offering significant potential for precision and efficacy, advancing these strategies requires addressing challenges such as incomplete AI data, regulatory uncertainties, and the lack of long-term studies. By overcoming these obstacles, clinicians can better meet individual health needs, prevent diseases, and integrate precision nutrition into routine care.
膳食补充剂（DS）被广泛用于解决营养缺乏问题和促进健康，但其滥用往往会导致疗效降低、产生不良反应以及引发安全问题。生物标志物驱动的方法已成为优化DS处方的一种有前景的策略，可确保精准性并降低与通用建议相关的风险。本叙述性综述综合了关于生物标志物指导的膳食补充以及人工智能（AI）在生物标志物分析中的整合的关键研究结果。对关键生物标志物类别——基因组学、蛋白质组学、代谢组学、脂质组学、微生物组学和免疫学——进行了综述，同时还介绍了用于解读这些生物标志物和定制补充剂处方的AI应用。生物标志物能够识别缺乏、代谢失衡和疾病易感性，支持有针对性且安全地使用DS。例如，像亚甲基四氢叶酸还原酶（MTHFR）基因多态性这样的基因组标志物可告知叶酸补充需求，而诸如葡萄糖和胰岛素水平等代谢组学标志物则指导对代谢紊乱的干预。AI驱动的工具通过考虑复杂的生物标志物相互作用和个体需求，简化了生物标志物的解读，优化了补充剂的选择，并提高了治疗效果。尽管取得了这些进展，AI工具仍面临重大挑战，包括依赖不完整的训练数据集以及临床验证算法数量有限。此外，目前大多数研究集中在临床人群，限制了对更健康人群的推广性。长期研究仍然稀缺，引发了关于生物标志物指导补充的持续疗效和安全性的疑问。监管上的模糊性进一步使补充剂的分类复杂化，尤其是当组合表现出类似药物的效果时。由AI增强的生物标志物指导的DS处方是个性化营养的基石。虽然具有显著的精准性和疗效潜力，但推进这些策略需要应对诸如AI数据不完整、监管不确定性以及缺乏长期研究等挑战。通过克服这些障碍，临床医生可以更好地满足个体健康需求、预防疾病，并将精准营养融入日常护理。
