Scholarlink

<p>Data quality directly influences artificial intelligence (AI) systems’ performance and practical application effects. Highthroughput screening (HTS) can significantly shorten the time to find optimal reaction conditions and provide consistent data support for data-driven scientific research. This study established an automated chemical synthesis platform (AutoCSP), which integrates substance dispensing, transfer, reaction, and analysis functions and a control program that coordinates these processes. The platform screened 432 organic reactions related to synthesizing the anticancer drug Sonidegib, demonstrating results comparable to manual operation. The random forest algorithm identified anomalies with 98.3% accuracy. More importantly, machine learning algorithms demonstrated that the data obtained from the automated platform offers significant advantages in ensuring experimental consistency and affirmed the platform’s capability to generate high-quality data that meets the needs of AI analysis. This paper also addressed challenges of the automated synthesis system, like unstable wireless communication and visual recognition, offering insights for future developments in smarter, more efficient, and stable chemical synthesis systems.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Production and evaluation of high-throughput reaction data from an automated chemical synthesis platform

  • 期刊论文

  • 发表日期：

    2025年5月26日

查看全文
  • Leiyun Zhong,
  • Yiming Xu,
  • Xinghai Li,
  • Peihao Cheng,
  • Shengyang Tao

摘要

Data quality directly influences artificial intelligence (AI) systems’ performance and practical application effects. Highthroughput screening (HTS) can significantly shorten the time to find optimal reaction conditions and provide consistent data support for data-driven scientific research. This study established an automated chemical synthesis platform (AutoCSP), which integrates substance dispensing, transfer, reaction, and analysis functions and a control program that coordinates these processes. The platform screened 432 organic reactions related to synthesizing the anticancer drug Sonidegib, demonstrating results comparable to manual operation. The random forest algorithm identified anomalies with 98.3% accuracy. More importantly, machine learning algorithms demonstrated that the data obtained from the automated platform offers significant advantages in ensuring experimental consistency and affirmed the platform’s capability to generate high-quality data that meets the needs of AI analysis. This paper also addressed challenges of the automated synthesis system, like unstable wireless communication and visual recognition, offering insights for future developments in smarter, more efficient, and stable chemical synthesis systems.