Unit:
Κατεύθυνση ΒιοπληροφορικήLibrary of the School of Science
Supervisors info:
Δρ. Βασιλική Α. Οικονομίδου, Επίκουρη Καθηγήτρια, Τμήμα Βιολογίας, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Δρ. Δημήτριος Ι. Στραβοπόδης, Αναπληρωτής Καθηγητής, Τμήμα Βιολογίας, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Δρ. Βασίλειος Κουβέλης, Επίκουρος Καθηγητής, Τμήμα Βιολογίας, Εθνικό και Καποδιστριακό Πανεπιστήμιο Αθηνών
Original Title:
Ανάλυση αποτελεσμάτων Omics με στόχο τον προσδιορισμό βιοδεικτών
Translated title:
Omics data analysis for biomarkers identification
Summary:
β-Thalassaemia is a hemoglobinopathy characterized by ineffective erythropoiesis. A large
number of mutations has been identified in the β-globin gene, resulting in reduced production of β-globin. The phenotype of a person with β-Thalassaemia can be asymptomatic, intermediate (Thalassaemia Intermedia – TI) or severe (Thalassaemia Major – TM). Each phenotype requires different therapeutic management (such as blood transfusion, iron chelation or bone marrow transplantation). Identification of molecular biomarkers for patient stratification is necessary to guide therapeutic decisions in clinical practice. In this study, gene expression data (RNA-Seq) from 2999 statistically significant differentially expressed genes from 49 individuals (Healthy, TI and TM) have been used to train a Random Forest machine learning tool named GeneSrf, to identify genes that distinguish different groups and might be used as biomarkers for patient stratification in thalassaemia. Our results show that RNA-Seq data does increase the success rate of the predictive model generated by the Random Forest algorithm to stratify TM vs H and TI vs H patients. However, stratification of TM vs TI patients was not possible. This study has identified genes that will be further experimentally validated and might be useful as biomarkers for efficient β-thalassaemia patient stratification.
Main subject category:
Science
Keywords:
β-thalassaemia, biomarker, stratification, classification, random forest, machine learning, biology, bioinformatics
