論文

査読有り 最終著者 国際誌
2020年4月8日

Large-scale single-molecule imaging aided by artificial intelligence.

Microscopy (Oxford, England)
  • Michio Hiroshima
  • ,
  • Masato Yasui
  • ,
  • Masahiro Ueda

69
2
開始ページ
69
終了ページ
78
記述言語
英語
掲載種別
研究論文(学術雑誌)
DOI
10.1093/jmicro/dfz116

Single-molecule imaging analysis has been applied to study the dynamics and kinetics of molecular behaviors and interactions in living cells. In spite of its high potential as a technique to investigate the molecular mechanisms of cellular phenomena, single-molecule imaging analysis has not been extended to a large scale of molecules in cells due to the low measurement throughput as well as required expertise. To overcome these problems, we have automated the imaging processes by using computer operations, robotics and artificial intelligence (AI). AI is an ideal substitute for expertise to obtain high-quality images for quantitative analysis. Our automated in-cell single-molecule imaging system, AiSIS, could analyze 1600 cells in 1 day, which corresponds to ∼ 100-fold higher efficiency than manual analysis. The large-scale analysis revealed cell-to-cell heterogeneity in the molecular behavior, which had not been recognized in previous studies. An analysis of the receptor behavior and downstream signaling was accomplished within a significantly reduced time frame and revealed the detailed activation scheme of signal transduction, advancing cell biology research. Furthermore, by combining the high-throughput analysis with our previous finding that a receptor changes its behavioral dynamics depending on the presence of a ligand/agonist or inhibitor/antagonist, we show that AiSIS is applicable to comprehensive pharmacological analysis such as drug screening. This AI-aided automation has wide applications for single-molecule analysis.

リンク情報
DOI
https://doi.org/10.1093/jmicro/dfz116
PubMed
https://www.ncbi.nlm.nih.gov/pubmed/32090254
ID情報
  • DOI : 10.1093/jmicro/dfz116
  • PubMed ID : 32090254

エクスポート
BibTeX RIS