Tomohiro Tobino

活性汚泥法の維持管理において行われる検鏡試験の自動化を目指した研究報告例はあるが, モデル構築のための教師画像として種別の微小動物画像を多数収集する必要がある。本研究は, 微小動物の網羅的な検出と外観に基づく分類を実施する画像解析モデルの開発を目的とし, 物体検出及び自己教師あり学習の知見を用いて, 微小動物検出器, 特徴量抽出器, 微小動物分類器の三段からなるモデルを構築し, 微小動物の検出・分類精度と本モデルによって得られた微小動物相と細菌叢との関連性を検討した。微小動物検出器ではAP50 = 82.85%の精度が得られ, 1種あたり15枚の画像から学習させた微小動物分類器では用意した誤検知画像の半分を除外できるモデルにおいて83.3%の正答率を得た。また, 12個の活性汚泥試料に本モデルを適用して得られた微小動物相構成とDNA シーケンシングにより得られた細菌叢構成との間に有意な関連性が示された。

A non-chemical solution is needed to control biofilms in water and wastewater treatment systems. High-frequency alternating electric field application offers an alternative approach that does not involve undesired electrode surface reactions. However, the effect of high-frequency alternating electric fields on bacterial cells in the attached-growth mode remains unexplored. This study investigated the impact of such fields on two stages of the biofilm development process: the initial adhesion phase (stage 1) and the early development phase (stage 2). Experiments were conducted using Escherichia coli and Pseudomonas aeruginosa in a three-channel flow cell exposed to alternating electric fields (3.05 V/cm, 20 MHz). The primary outcome of this study demonstrated that alternating electric fields decreased adhered cell numbers at both stages due to their inhibitory effect on growth. The alternating electric fields also triggered cell detachment after the initial attachment stage but not in mature biofilms. Interestingly, despite a reduction in cell counts, the amount of total biofilm biomass remained unaffected, which was likely due to increased cell size via cell elongation compensating for the decrease in numbers. No synergistic effects with respect to hydrodynamic forces were observed. These findings highlight the potential applicability of alternating electric fields to biofilm control and provide implications for water and wastewater engineering applications.

The use of randomly generated DNA fragment sequences as probes on DNA arrays offers a unique potential for exploring unsequenced microorganisms. In t