MLb-LDLr: LDLaren hartzaile aldaeren eragina aurresateko ikasketa automatikoko eredua

Abstract

Hiperkolesterolemia Familiarrak plasmako dentsitate baxuko lipoproteina-kolesterol maila altua du ezaugarri nagusi. Hiperkolesterolemia Familiarra agertzearen arrazoirik ohikoena LDL hartzailearen aldaerak dira, zeintzuek dentsitate baxuko lipoproteina-kolesterola hartzea murrizten eta plasman pilatzea eragiten duten. Gaixotasunaren diagnosi goiztiarra gakoa da gaixotasun kardiobaskularraren agerpena saihesteko. Alderen karakterizazioan gehien erabilitako tekniken prozesuak luzeak eta garestiak dira. In silico softwareak, aldiz, non algoritmo konputazionalak erabiltzen diren aldaera baten eragina aurresateko, konponbide berritzaile bat dira. Lan honetan LDL hartzailearen missense aldaeren patogenizitatea aurresan dezakeen ikasketa automatikoko eredu bat garatu dugu.Familial hypercholesterolemia is characterized by high concentration of low-density lipoprotein cholesterol in plasma. The most common cause of Familial hypercholesterolemia are mutations in LDL receptor (LDLr) gene, which reduce LDL-C uptake and promote the accumulation of LDL-C in plasma. An early diagnosis of the disease is key to avoiding cardiovascular disease development. The most used characterization techniques are long and expensive processes. On the other hand, in silico software, where computational algorithms are used to predict the effect of a variant, is an innovative solution. In this work, we have developed a Machine Learning model that can predict the pathogenicity of LDLr missense variants.