.png){kind=icon}
.png){kind=icon}
Return
Machine Learning-Assisted Quantification of Organelle Abundance
.png){kind=icon}
PRE
AI
A
C
N
R
M
S
H
T
DOI：10.21769/BioProtoc.5626.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Organelle abundance is a key microscopic readout of organelle formation and, in many cases, function. Quantification of organelle abundance using confocal microscopy requires estimating their area based on the fluorescence intensity of compartment-specific markers. This analysis usually depends on a user-defined intensity threshold to distinguish organelle regions from the surrounding cytoplasm, which introduces potential bias and variability. To address this issue, we present a machine learning-assisted algorithm that allows for the quantification of organelle density using the open-source Fiji platform and WEKA segmentation. Our method enables the automated quantification of organelle number, area, and density by learning from training data. This standardizes threshold selection and minimizes user intervention. We demonstrate the utility of this approach for both membrane and non-membrane organelles, such as peroxisomes, lipid droplets, and stress granules, in human cells and whole fish samples.
Keywords:
Organelle
Peroxisome
Stress granule
Quantification
Abundance
Density
Machine learning
Fiji
WEKA segmentation
PeroxiSPY
Journal
B
IF:
1.1
Papers:
191
Citations:
5.4K
Organization
Cited Papers