An Overview of Artificial Intelligence-based Extracellular Vesicle Characterization Methods
محل انتشار: اولین کنگره بین المللی هوش مصنوعی در علوم پزشکی
سال انتشار: 1402
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 156
نسخه کامل این مقاله ارائه نشده است و در دسترس نمی باشد
- صدور گواهی نمایه سازی
- من نویسنده این مقاله هستم
استخراج به نرم افزارهای پژوهشی:
شناسه ملی سند علمی:
AIMS01_126
تاریخ نمایه سازی: 1 مرداد 1402
چکیده مقاله:
Background and aims: Extracellular Vesicles (EVs) are cell membrane-derived nano-particlesthat can be secreted by all cells. These particles have been described as double-edged swordsbecause of their crucial role in a wide range of physiological and pathological conditions. SmallEVs (Exosomes), as one of the most important types of extracellular vesicles, are particles with adiameter of ۳۰ to ۲۰۰ nm that can carry a variety of large macromolecules depending on the typeof their mother cell. Several studies have shown that by examining the composition of small EVsin body fluids, it is possible to diagnose pathological conditions before the appearance of clinicalsymptoms. Hence, small EVs are considered promising biomarkers for the early detection ofdiseases. Despite the great diagnostic and therapeutic potential of these nanoparticles, the heterogeneityof small EVs makes it difficult to analyze their cargo and structure. Artificial intelligence(AI) and machine learning can be helpful in this field.Method: This study is a review of published articles in the field of AI-based extracellular vesiclecharacterization methods. In order to collect the articles, the keywords of “Extracellular vesicles”,“Exosomes”, “Artificial intelligence”, and “Machine learning” were searched in databases suchas Google Scholar, Science Direct, PubMed and etc.Results: The results of this review demonstrated that most of the published studies in this field canbe categorized into two major groups: ۱- utilizing AI-based methods to investigate the alterationsof exosomal microRNA expression patterns in different physiological and pathological conditions.For instance, supervised machine learning algorithms have shown that Lyssavirus-infectedhuman neurons have a unique cellular and exosomal miRNA signature. In this regard, supervisedmachine learning algorithms determined ۶ cellular miRNAs (miR-۹۹b-۵p, miR-۳۴۶, miR-۵۷۰۱,miR-۱۳۸-۲-۳p, miR-۶۵۱-۵p, and miR-۷۹۷۷) were indicative of lyssavirus infection. This findingcan be used as a diagnostic biomarker for the early detection of rabies. Using a similar strategy,specific exosomal miRNA profiles have been discovered for many tumor types such as pancreaticand breast tumors. ۲- Using AI-based image processing methods to investigate the morphologicalchanges of exosomes. Recent publications have shown the morphological changes of exosomesin various pathological conditions. For Instance, the results of several recently published studieshave shown that exosomes secreted from cancer cells have a larger average diameter comparedto the exosomes secreted from normal body cells. Identifying these changes by using AI-basedmethods can play an important role in the early diagnosis of diseases.Conclusion: Utilizing AI-based analytical methods can greatly help to expand the diagnostic andtherapeutic potentials of EVs.
نویسندگان
Pouya Houshmand
Department of Microbiology and Immunology, Faculty of Veterinary Medicine, University of Tehran, Tehran, Iran
Parham Soufizadeh
Gene Therapy Research Center, Digestive Diseases Research Institute, Shariati Hospital, Tehran University of Medical Sciences, Tehran, Iran
Ali Pourshaban-Shahrestani
Student of Veterinary Medicine, Faculty of Veterinary Medicine, University of Tehran, Tehran, Iran
Gholamreza Nikbakht Brujeni
Department of Microbiology and Immunology, Faculty of Veterinary Medicine, University of Tehran, Tehran, Iran