皮膚細胞粒線體三維影像分析系統

Three-Dimensional Image Analysis System for Mitochondrial of Skin Cell

呂珈玫(Jia-Mei Lu)

指導教授 ： 蔡育秀

中原大學/工學院/生物醫學工程研究所/碩士(2016年)

https://doi.org/10.6840/cycu201600276

摘要

本研究針對不同層早衰細胞粒線體與正常細胞粒線體進行影像處理和分析，並比較早衰與正常細胞粒線體的形態差異。首先在每一張同一個視野中圈選單一細胞粒線體，並將細胞粒線體分成Layer1與Layer2兩層。以MicroP進行影像分割，使二值化影像利於粒線體三維特徵的擷取。為了區分粒線體的次形態，將粒線體三維形態特徵做正規化，使得所有的特徵差異變小，再使用X-Means群集分析，依照子集與子集中心距離排列，並定義了7種形態，球狀，小塊狀、短條狀、大塊狀、短分支狀、中分支狀和長分支狀。再計算不同層早衰與正常細胞粒線體數量、體積、平均形態、分支與整體細胞粒線體的分布情況等等的特徵。最後使用J48決策樹，將所有細胞特徵資訊以70%訓練組與30%測試組的細胞分類。
統計分析顯示，第一層早衰細胞粒線體的粒線體數量、體積與粒線體在細胞中的分布情形，皆較於正常細胞粒線體與第二層早衰粒線體多且分布廣，在形態分析中，第一層早衰細胞粒線體的球狀、小塊狀和短條狀平均體積較多，第一層正常細胞粒線體多為大塊狀與短分支狀，第二層早衰與正常粒線體有相近的細胞特徵。訓練組及測試組正確率分別為95.12%與82.35%。

關鍵字

粒線體三維形態、三維影像分析、CISD2、早衰

並列摘要

This study is focused on effects of CISD2 knock out on three dimension mitochondrial morphology, which differ epidermal Layers from skin tissues in mice. Skin mitochondrial micrographs are segmented into single-cell of three dimension micrographs by imageJ. According to depth of skin tissues, single-cell micrographs are grouped in to Layers1 and 2, there is wild type to compare with CISD2 knock out. Segment the images to be the binary by MicroP. Extracted morphological features of individual mitochondria from every single cell are applied to classify mitochondrial morphology by use of X-means clustering. There are 7 mitochondrial morphological subtypes identified, including small globules, lumps, large lumps, short tubules, short small branched tubules, medium branched tubules and long branched tubules. Features of mitochondrial morphology are used to train a classifier to differentiate mitochondrial morphology in different Layers of skin tissues of CISD2 KO and wild type mice by use of decision tree.
The classifier can distinguish the effects of CISD2 KO on mitochondrial morphology with accuracy 97% in training and 82% in testing. From Statistics, The total object number of CISD2 KO Layer1 mitochondrial is more than others. The depth of wild type Layer2 mitochondrial is deeper. CISD2 KO results in morphology changed in Layer1 and Layer2.

並列關鍵字

Mitochondrial morphology, numerical 3D image analysis, CISD2, premature ageing

參考文獻

[12] 洪婉綺, “白血球粒線體形態變化之三維影像分析系統,” 生物醫學工程學系, 中原大學, 2013.

[15] 林俊志, “自動化粒線體之型態分析系統,” 生物醫學工程學系, 中原大學, 2010.

[1] R. K. Chaturvedi and M. F. Beal, “Mitochondrial diseases of the brain,” Free Radic. Biol. Med., vol. 63, pp. 1–29, Oct. 2013.

[4] S. C. Correia, R. X. Santos, G. Perry, X. Zhu, P. I. Moreira, and M. A. Smith, “Mitochondrial Importance in Alzheimer’s, Huntington's and Parkinson's Diseases,” in Neurodegenerative Diseases, vol. 724, 2012, pp. 205–221.

[5] L. C. Gomes and L. Scorrano, “Mitochondrial morphology in mitophagy and macroautophagy,” Biochim. Biophys. Acta - Mol. Cell Res., vol. 1833, no. 1, pp. 205–212, Jan. 2013.