Ivan Villalba Electrical Engineering Cal Poly SLO Oxnard College - - PowerPoint PPT Presentation

▶

Mar 08, 2023 155 likes •740 views

Vision Research Lab Center for Bio-image Informatics Ivan Villalba Electrical Engineering Cal Poly SLO Oxnard College (2005) Mentors Jiyun Byun DeeAnn Hartung Faculty Advisor Dr. B.S. Manjunath Introduction Alzheimers

SLIDE 1

Mentors Jiyun Byun DeeAnn Hartung

Ivan Villalba

Vision Research Lab Center for Bio-image Informatics

Faculty Advisor

Dr. B.S. Manjunath

Electrical Engineering Cal Poly SLO Oxnard College (2005)

SLIDE 2

Introduction

– Alzheimer’s

Image Analysis

– Properties – Procedure

Experimental Results

– Cell Segmentation – Cell Density

Future Work

SLIDE 3

SLIDE 4

Characterized by two pathological hallmarks in the brain,

1. Extracellular Amyloid Plagues

Amyloid-Beta (Aβ) protein

2. Intracellular Neurofibrillary Tangles (NFTs)

Tau protein

Aβ ? Tau Alzheimer’s

Cell Death Dysfunction Dysfunction

Big Picture: Aβ and Tau

– Examining molecular mechanism of cell death.

SLIDE 5

Model System

– Monkey Kidney Cells – Cancer cell line – COS1 Cells transfected with Tau

Determine cell survival/death under

various treatments and time points for Alzheimer’s disease.

SLIDE 6

Cell survival/death ratio
Cell Count by segmentation

– Live cells = total cells – dead cells

Cell shape: extract further information

(e.g. Cell division ratio)

SLIDE 7

2 Hours 120 Hours Various Treatments

SLIDE 8

SLIDE 9

10x Confocal microscope 1024x1024 pix 1 pix = 0.1243µm

SLIDE 10

– Total Cells – Dead Cells

(Enters leaky membranes)

– Live Cells

(Hydrolyzed by intracellular enzymes)

Hoechst Channel

SLIDE 11

Isolate Channels

Hoechst Channel

SLIDE 12

Isolate Channels
Median Filtering

– Noise Reduction

Hoechst Channel

SLIDE 13

Isolate Channels
Median Filtering
Binary Image

– Threshold

Hoechst Channel

SLIDE 14

Isolate Channels
Median Filtering
Binary Image
Label Image

Hoechst Channel

SLIDE 15

Isolate Channels
Median Filtering
Binary Image
Label Image

Example: 10x8 pixel Binary image

SLIDE 16

Isolate Channels
Median Filtering
Binary Image
Label Image

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1

Example: 10x8 pixel Label Image

SLIDE 17

Isolate Channels
Median Filtering
Binary Image
Label Image

Hoechst Channel

SLIDE 18

Isolate Channels
Median Filtering
Binary Image
Label Image
Object Analysis

SLIDE 19

Isolate Channels
Median Filtering
Binary Image
Label Image
Object Analysis

120 Hours Untreated

SLIDE 20

Isolate Channels
Median Filtering
Binary Image
Label Image
Object Analysis

– Normal (1 cell)

120 Hours Untreated

SLIDE 21

Isolate Channels
Median Filtering
Binary Image
Label Image
Object Analysis

– Cluster ( >= 3 cells)

120 Hours Untreated

SLIDE 22

Object Analysis

– Normal – Peanut

(Dividing cell)

– Cluster ( >= 3 cells)

SLIDE 23

Area

9 1 8

SLIDE 24

Area
Perimeter

1 1 7

SLIDE 25

Area
Perimeter
Compactness

Compactness = 4 x pi x Area Perimeter2

Normal

Compactness = 1

Low Compactness Cluster

SLIDE 26

Area
Perimeter
Compactness
Minor Axis

5 3

SLIDE 27

Area
Perimeter
Compactness
Minor Axis
Major Axis

6 4

SLIDE 28

Area
Perimeter
Compactness
Minor Axis
Major Axis
Axis Ratio

Axis Ratio = Minor Axis Major Axis

SLIDE 29

2hrs Untreated

Class Type:

1. Normal + Peanut
2. Cluster

SLIDE 30

Crop Regions

2hrs Untreated

SLIDE 31

Crop Images

2hrs Untreated

SLIDE 32

Area
Perimeter
Compactness
Minor Axis
Major Axis
Axis Ratio
Clusters can be discriminated.
Normal & Peanut are difficult to discriminate.

SLIDE 33

SLIDE 34

42 “Untreated” Images from 7 different time

points

2 hrs (7)
6 hrs (5)
12 hrs (6)
24 hrs (5)
48 hrs (8)
72 hrs (6)
120 hrs (5)

SLIDE 35

Red = Normal + Peanut Green = Clusters 120 Hours

SLIDE 36

Red = Normal + Peanut Green = Clusters 24 Hours

SLIDE 37

Cell Density:

– Cells = Number of “Normal & Peanut” – Area = Image Area – Cluster Area

Cell Density = Cells Area

SLIDE 38

Cell Density (Untreated)

SLIDE 39

SLIDE 40

Apply to various treatment

– Untreated – 5 uM Staurosporine (stauro) – 10 uM Amyloid-Beta () – 100 nM Taxol (Taxol) – 100 nM Taxol + 10 uM Amyloid-Beta (Tx-)

Cell survival/death ratio.

SLIDE 41

Better Threshold

– Grayscale Image Binary Image

Improve Segmentation

SLIDE 42

2hrs Hoechst Channel Red = Normal Green = Peanut

SLIDE 43

SLIDE 44

SLIDE 45

SLIDE 46

Faculty Advisor

– Dr. B.S. Manjunath

Mentors:

– Jiyun Byun – DeeAnn Hartung

Center for Bio-Image Informatics

SLIDE 47

SLIDE 48

Characterized by two pathological hallmarks in the brain,

1. Extracellular Amyloid Plagues

Amyloid-Beta (Aβ) protein

2. Intracellular Neurofibrillary Tangles (NFTs)

Tau protein

Aβ ? Tau Alzheimer’s

Cell Death Dysfunction Dysfunction

SLIDE 49

Amyloid-Beta (Aβ)

1. Amyloid Percursor Protein (APP)
a. Function: Undetermined
2. Can’t cause disease without Tau

Aβ ? Tau Alzheimer’s

Cell Death Dysfunction Dysfunction

SLIDE 50

Amyloid-Beta (Aβ)

SLIDE 51

Amyloid-Beta (Aβ)

SLIDE 52

Tau Protein

1. Function:

Regulates neuronal microtubule dynamics

SLIDE 53

Aβ ? Tau Alzheimer’s

Cell Death Dysfunction Dysfunction

SLIDE 54

Experimental Setup

Treatments

– Untreated – 5 uM Staurosporine – 10 uM Aβ – 100 nM Taxol – 100nM Taxol + 10 uM Aβ

Timepoints

– 2 hours – 6 hours – 12 hours – 24 hours – 48 hours – 72 hours – 120 hours

SLIDE 55

Threshold

Logical Image

For Original Image

Pixels Pixel Value

SLIDE 56

Threshold

Logical Image

For Filtered Image

Pixels Pixel Value