請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86103
標題: | 整合類神經系統及反應曲面法探討牛樟芝液態栽培最適化條件 A Hybrid of Response Surface Methodology and Artificial Neural Network in the Optimization of culture conditions for mycelia growth of Antrodia cinnamomea |
作者: | Meng-Hsin Lee 李孟信 |
指導教授: | 張斐章(Fi-John Chang) 張斐章(Fi-John Chang | changfj@ntu.edu.tw | ), |
關鍵字: | 牛樟芝,反應曲面法,生物質量,類神經網路,菌絲體生長, Antrodia cinnamomea,Biomass production,Mycelia growth,Response surface methodology,RSM,ANN,Artificial neural network, |
出版年 : | 2022 |
學位: | 博士 |
摘要: | 牛樟芝 (Antrodia cinnamomea) 如要發展成為保健產品和功能性商品時,需面臨大量生產及品質一致的挑戰。隨著自然環境的永續發展,實迫切需要以實驗室培養系統增加牛樟芝的產量,以彌補其產量下降和野外棲息地破壞的事實。由於影響微生物生物質產量(菌絲產量)的培養條件有多種因素組合,因此為了建立牛樟芝菌絲產量的最佳條件是一個繁瑣的過程。本研究目的使用混合類神經網路(ANN)法和反應曲面法(RSM)於建立「增加牛樟芝菌絲體產量、生長效率及產量效率的最佳條件」。反應曲面法目的在優化培養條件,而類神經網路目的在精準預測生物質量生產的主導因素,因而降低實驗次數及節省時間和成本。本次選擇了七個主要影響因子來評估這七個因子的最佳濃度和條件對生物質生產的影響,包括葡萄糖、馬鈴薯泥培養基(PDB)、樟木(CC)-多醣(PS)、樟腦、CC水提取物、洋菜和不同培養基起始pH值的培養條件。在反應曲面實驗階段應用Plackett-Burman及二水準部分因子實驗( 27-2 ) ,確定了四個關鍵因子。再以四因子六個水準的中心混合設計,來研究生物質量與關鍵因素之間的相關性。結果顯示,最終培養牛樟芝菌種B85的最佳培養基組成為葡萄糖20.83 g/L、PDB 40.44 g/L、CC-PS 0.16 g/L、樟腦1 g/L、CC 水提取物20 g/L、洋菜3.47 g/L和pH 7.37。反應曲面法所得的生物質量比對照組的生物質量高200 %。類神經網路的加入,透過機器學習技術進行最佳條件預測,能夠縮短反應曲面法實驗次數。本研究所提出的方法可以在實驗室培養的最佳條件下提供可靠的藥用真菌生產,與傳統的窮舉法(所有的條件組合)實驗程序相比,可減少成本、時間和精力。在此最佳生長條件下,於工廠大量生產所需的牛樟芝,保護野生牛樟樹及牛樟芝,保有於自然界中緩慢生長的特性,還可達到聯合國17項永續發展目標中的第12項目標:確保永續的消費和生產模式(SDG 12);及第15項目標:保育和永續利用陸域生態系統,永續管理森林,防治沙漠化,防止土地劣化,遏止生物多樣性的喪失(SDG 15)。此外,本研究提出一個以混合RSM-ANN深具潛力的方法,在未來遇到多因子試驗任務時,希望能夠帶來新一代的生物質量生產技術。 Antrodia cinnamomea has faced the challenge to cope with mass production for functional food and nutraceutical development. With the distinctive habitat, there is an urgent need to increase the production of A.cinnamomea through laboratory cultivation to make up for its declining yield and habitat destruction in the wild. Since mycelial biomass (mycelia yield) production is affected by a combination of various factors, it is a tedious process to obtain the optimal condition for production A. cinnamomea mycelia. The aim of this study was to enhance the biomass production of A. cinnamomea by a methodology that hybrids Artificial Neural Network (ANN) and Response Surface Methodology (RSM). RMS was performed to optimize the cultural condition, and ANN was used to discover the dominating factors of biomass production. Seven major affecting factors including agar, glucose, Cinnamomea camphora (CC)-polysaccharide (PS), camphor, CC water extract, PDB, and different pH of culture medium, were chosen to evaluate the impacts of optimal concentrations and condition on biomass production. The 32 (27-2) fractional factorial designs and Plackett-Burman design identified four major factors. A four-factor six-level central composite design was performed to study the correlation between biomass and the major factors. The results showed that the final optimal medium compositions for the cultivation of A. cinnamomea B85 were be 3.47 g/L of agar, 20.83 g/L of glucose, 0.16 g/L of CC-PS, 1 g/L of camphor, 20 g/L of CC water extract, 40.44 g/L of PDB, and 7.37 for pH of the culture medium. The yield of RSM was 200% higher than that of the control. The proposed RSM-ANN can apply to construct different rules during the learning process so that the interdependence between seven input factors and mycelia yield can be explained which allows us to discover how the model obtains a satisfactory prediction. In this process, four key factors for improved mycelia yield can be identified according to the rules of RSM-ANN. The proposed methodology can offer a mass production of the entitled fungus under optimum conditions, which have the advamtages of a reduction in the time, effort, and cost compared to the traditional labor-intensive experimental procedures, while also meeting the Sustainable Development Goals (i.e. SDG 12 and SDG 15), hitting a new milestone. ANN offers a new opportunity in the prediction of mycelial growth systems. Also, we propose the potential of a hybrid RSM-ANN method for solving the difficulties of multifarious tasks in the future with the hope to create a new generation of biomass production technologies. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86103 |
DOI: | 10.6342/NTU202203246 |
全文授權: | 同意授權(全球公開) |
電子全文公開日期: | 2022-09-13 |
顯示於系所單位: | 生物環境系統工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-0809202211015300.pdf | 7.51 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。