ANN data analysis of complex interactions in SALS: a multidisciplinary study

Amyotrophic Lateral Sclerosis (ALS) is a rare neurodegenerative disorder characterized by selective degeneration of both upper and lower motor neurons in the brain, brainstem, and spinal cord. This results in paralysis due to muscle weakness and atrophy, leading to death in 3-5 years. Artificial neural network (ANN) analyses shift the paradigm of data analysis into a bottom-up definition of the regulating processes, generated on the basis of the relations between data. Auto Contractive Map (Auto-CM, based on ANN architecture) allows to visualize the correlations among the variables by constructing a space where closeness reflects accurately their association. Patients and controls originating from a small and defined geographical area were recruited to limit environmental exposure variability. Data collected from a food frequency questionnaire were analysed with this approach, excluding evident differences in the nutrition habits of ALS subjects with respect to healthy controls. Variables best describing the disease status were: fruit, vegetables, white meat consumption and a lower BMI, whereas sweets, bread and croissant consumption together with a higher BMI were associated to the control group. A metallomic evaluation of sera was performed with ICP-MS, leading to the association of low levels of As and a metal overload potentially originating from tap water consumption with the ALS group. Auto-CM analysis of proteomic data obtained with 2DE in non-reducing conditions associated low levels of apolipoproteins and high levels of zinc-alpha-2-glycoprotein and clusterin with ALS group, highlighting deregulations in lipid metabolism. Results show that the application of new statistical analyses, based on machine learning, would be the key to translate complex raw data into understandable models by integrating different fields of knowledge, to create a picture of the several factors involved in the disease aetiology and course, complete as much as possible