Rapid detection of red cell membrane defects leading to hemolytic anaemias

Abstract

Hemolytic anaemias are a group of disorders characterised by the loss of integrity of the red blood cell membrane that leads to premature RBC clearance. These conditions often are heterogeneous in the genetic causes, complicating diagnosis by high throughput DNA sequencing. We applied deep learning technologies to build a diagnostic tool for hemolytic anaemias. We used an Imaging Flow Cytometer to obtain images of red blood cell membranes for several hemolytic anaemias and then trained the deep neural network to distinguish the stages of the disease using Keras and TensorFlow. This project combines Python-based machine learning with socially viable healthcare applications.