Machine learning algorithm selects transcripts in blood to distinguish Alzheimer’s disease and other neurodegenerative diseases

Abstract Background Clinical diagnosis of neurodegenerative diseases is notoriously inaccurate and current methods are often expensive, time‐consuming, or invasive. Simple inexpensive and noninvasive methods of diagnosis could provide valuable support for clinicians when combined with cognitive assessment scores thereby reducing health care costs by ruling out or streamlining additional tests such as imaging. Biological processes leading to neuropathology are reflected in both the central nervous system and vascular peripheral system and progress silently for many years. Because Alzheimer’s disease (AD) symptoms often overlap with common disorders that may be treatable or reversible, an accurate diagnosis at the earliest time is crucial. A blood‐based screen to distinguish and classify neurodegenerative diseases is especially interesting having low cost, minimal invasiveness, and accessibility to almost any world clinic. Additionally, an inexpensive blood panel could also monitor disease progression, for example, marking the advancement beyond mild cognitive impairment (MCI). Method The study of blood‐based changes in mRNA gene expression presents a good strategy for differentiating patients of any neurodegenerative disease regardless of the proteins or their post‐translational modifications occurring in disease. Information about features selected for the classification may also guide insight into possible treatment strategies. Using existing public datasets, we developed a machine learning algorithm for application on transcripts present in blood. Result Our machine learning algorithm generated small sets of transcripts which distinguish a number of neurodegenerative diseases with high sensitivity and specificity. Conclusion Neurodegenerative diseases such as AD are associated with changes in specific molecular pathways and signatures. Using RNA expression as blood biomarkers can also provide information for a pathophysiological relationship with disease and gives us the advantage of using the vast knowledge of gene expression to not only predict disease but also analyze the pathways involved. Our chosen transcripts reveal that neurodegenerative diseases have common themes which after removal bare the unique transcripts of each disease.