Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Genome-wide microscopic image analysis of potential regulators of lysosomes and cell death

Miari, Mariam (2021) BINP51 20201
Degree Projects in Bioinformatics
Popular Abstract
When technology meets cell biology

Lysosomes are small bags present in our cells that are filled with destructive molecules, normally needed for removing cell garbage. In some cases, these bags rupture and their contents leak into the interior of the cell leading to cell death. This can contribute to many diseases, such as Alzheimer’s disease. However, rupturing lysosomes intentionally, can also be a strategy to target cancer cells! To understand lysosomes and cell death better and develop therapies that target them, we need to be able to test how thousands of therapeutic drugs and genetic changes affect them. Ruptured lysosomes, and many other structures important for cell death, like the cell nucleus, can be observed in a microscope... (More)
When technology meets cell biology

Lysosomes are small bags present in our cells that are filled with destructive molecules, normally needed for removing cell garbage. In some cases, these bags rupture and their contents leak into the interior of the cell leading to cell death. This can contribute to many diseases, such as Alzheimer’s disease. However, rupturing lysosomes intentionally, can also be a strategy to target cancer cells! To understand lysosomes and cell death better and develop therapies that target them, we need to be able to test how thousands of therapeutic drugs and genetic changes affect them. Ruptured lysosomes, and many other structures important for cell death, like the cell nucleus, can be observed in a microscope but measuring them by hand is too time-consuming. In our project, we used the computer to analyze microscopic images to automatically identify bone cancer cells and structures inside the cells, such as nuclei and lysosomes. Besides, we assessed the possibility of training a computer program to distinguish images of healthy cells from the ones with ruptured lysosomes

Before visualizing the cells under the microscope, cells had been dyed along with their subcellular structures with specific dyes. Once the desired structures became visible, images had been captured in a microscope and we analyzed these images with a software, called CellProfiler. The software helped us to identify the target structures, to count them, and to know their sizes and shapes as well as their brightness patterns. Since CellProfiler needs a lot of human intervention to find the best settings, we also compared it to another method that is based on artificial intelligence, which learns on its own from example images where the cell structures have been outlined manually. Lastly, we wanted the computer to distinguish between healthy cells and those with ruptured lysosomes, so we ‘trained’ another artificial intelligence model to do so, i.e. we provided it with a large number of images of each type so that it can learn to extract m meaningful patterns that distinguish them.

We found that the computer performed very well in the tasks that humans used to do manually in the past. Both methods identified the cells and structures inside them quickly and efficiently. This result is very important, because it means that spending so much time and effort on studying cells images and on counting the cells and subcellular structures is not necessary anymore. We found that the artificial intelligence method, which is very self-dependent, generally outperformed CellProfiler, which still relies on the researcher in almost all the steps of analysis. Our second main result showed that the trained artificial intelligence model was able to distinguish images of healthy cells from the ones with ruptured lysosomes (i.e. probably dying cells) with very high accuracy

Master’s Degree Project in Biology/Molecular Biology/Bioinformatics 45crs.
Department of Biology, Lund University
Advisor: Sonja Aits
Cell death, lysosomes, and artificial intelligence group/ Biomedical Centre (BMC) (Less)
Please use this url to cite or link to this publication:
author
Miari, Mariam
supervisor
organization
course
BINP51 20201
year
type
H2 - Master's Degree (Two Years)
subject
language
English
id
9066762
date added to LUP
2021-10-12 15:15:52
date last changed
2021-10-12 15:15:52
@misc{9066762,
  author       = {{Miari, Mariam}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Genome-wide microscopic image analysis of potential regulators of lysosomes and cell death}},
  year         = {{2021}},
}