螢光蛋白質自動辨識及追蹤系統

An Automated Fluorescent-tagged Protein Tracking System

黃芸妮(Yun-Ni Huang)

指導教授 ： 蔡育秀 ； 林崇智

中原大學/工學院/醫學工程研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400682

摘要

摘要

蛋白質的動態分布情形會顯露該蛋白質功能和生理機制，例如網狀活化系統活化時，蛋白質會由細胞質移轉至細胞膜，而大多數的蛋白質是透過胞吐方式運送、細胞膜的擴散作用和緊密結合細胞膜進而影響與之接觸的細胞。所以對於細胞而言，蛋白質動態變化情形是十分重要的生理參數，蛋白質的出現和消失對細胞的變化都是有意義的。在活細胞中觀察蛋白質的變化情形，是一項新的實驗模式和技術，本研究目的在建構自動辨識追蹤螢光蛋白質影像之應用系統，觀察蛋白質的動態分布與其功能和生理機制，為輔助細胞實驗的一套後端分析軟體，以追蹤蛋白質的變化為主，並將結果以數據和圖表的方式呈現，以減少研究人員人工觀察的時間和誤差，藉由分析軟體的驗證，研究人員可以將實驗結果數據化並做有利的說明和解釋。
由於細胞中的蛋白質影像有不同的變化情形，其中主要可分成亮度變化和位置變化。亮度變化有逐漸變暗、保持不變和逐漸上升三種狀況。位置上的變化比較不規則，大致可分成直線移動，波浪狀的擾動和原地旋轉等。根據這些情形，本研究設定各種模擬情境以測試所建構的系統。
研究結果顯示，對於蛋白質和背景的灰階值差異在10時，系統可以正確的辨識出蛋白質影像。研究測試了多組的模擬影像發現，當蛋白質在frame及frame之間移動達到10 pixels時，本系統仍可追蹤其到其移動軌跡。相較於所觀察的蛋白質在frame及frame之間平均移動速度約為5 pixels，本系統足以偵測蛋白質位置變化的各種情形。若蛋白質亮度逐漸增加則系統有100％的辨識追蹤效果。對於蛋白質的亮度逐漸變暗且下降幅度小於灰階值15時，系統可以很精確的追蹤該蛋白質直到其消失，勝過眼睛的直接辨識效果。在實際上蛋白質亮度變化起伏，配合多樣的顯示方式，有彩色的線條輔助和動態變化，使用者有較佳的理解性。

關鍵字

物體追蹤 ； 綠色螢光蛋白 ； 影像處理

並列摘要

ABSTRACT

Cells could recognize and combine with their message molecules only through objective protein. Objective protein has singularity that only certain cell message molecules could trigger it. To understand the links between cell message molecules and objective proteins is the key to explain the complex life phenomenon. This research has setup automated protein recognition and tracking system to support aforementioned proteomics related study. By utilizing image processing techniques, the motility of objective protein in a cell, recorded from a fluorescent microscope at consecutive time frames, can be derived. Current progress demonstrated the objective protein can be automatically distinguished from vesicles of cells. The speed and direction of protein movement can be traced and displayed in 3-D plots. An accompanied user-friendly graphical interface provided by the proposed system, has further reduced the complexity of operating flow.
To identify the system performance, different test scenarios have been constructed. These scenes simulated the proteins moving, hovering, and fusing in cells. To test robustness of the system, each test scene has dithered with noise and reduced the brightness difference to the background.
The results shown that the system can identify the proteins, if the gray level difference between proteins and background are more than 10. It is also found that even the protein’s position change between frames is up to 10 pixels, the system still can trace correctly trace the protein. Since, by observing the fluorescent images sequences, the average protein position change is around 5 pixels between consecutive frames, the system is capable to catch the different variety of protein movement. When the gray level of tracked object is monotonic increasing, the system has no error in detecting the object movement. For those proteins with decreasing gray level, the system can correctly track the movement until the gray level below 15. The system performed better than visual inspection. The variety presentation tools provided comprehensive ways to interpret the protein motility in the cells.

並列關鍵字

Image Processing ； Object Tracking ； Green Fluorescent Protein

參考文獻

[1]Mia K. Markey, Michael V. Boland, and Robert F. Murphy. “Toward Objective Selection of Representative Microscope Images”. Biophysical Journal. 76（1999）：2230-2237.

[2]Jianpixelng Fana and AhmedK. Elmagarmid. “An automatic algorithm for semantic object generation andtemporal tracking”. Signal Processing Image Communication 17(2002)：145-164.

[3]Seonghoon Kang, and Seong-Whan Lee. “Real-time tracking of multiple objects in space-variant vision based on magnocellular visual pathway”. Pattern Recognition 35 (2002)：2031-2040.

[4]Michael V. Boland and Robert F. Murphy. “A neural network classifier capable of recognizing the patterns of all major subcellular structures in fluorescence microscope images of HeLa cells”. BIOINFORMATICS 17(2001)：1213-1223.

[5]W. Tvarusko, M. Bentele, T. Misteli, R.Rudolf, C.Kaether, D. L. Spector, H. H. Gerdes, and R. Eils. “Time-resolved analysis and visualization of dynamic processes in living cell”. Cell Biology 96（1999）：7950–7955.