基于机器学习的藻类对微塑料和有机污染物的联合毒性的预测
Jing Lu1, Yang Song1, Bin Hou1
1School of Environment and Safety Engineering, North University of China, Taiyuan, 030051, China.
Marine pollution bulletin
|February 28, 2026
概括
机器学习模型有效地预测水生生态系统中的微塑料 (MP) 和有机污染物的联合毒性. 关键因素包括MP特性,污染物特性和实验条件,有助于风险评估.
科学领域:
- 环境化学环境化学
- 生态毒理学 生态毒理学
- 计算毒理学计算毒理学
背景情况:
- 水生生态系统面临微塑料 (MP) -有机污染物混合物的复杂毒性.
- 传统模型很难有效地捕捉这些关节毒性机制.
研究的目的:
- 开发MP-有机污染物混合物引起的藻类生长抑制的预测模型.
- 通过机器学习和SHAP分析,确定驱动关节毒性的关键因素和机制.
主要方法:
- 从现有文献中编制了10个MP和6种有机污染物的数据集.
- 开发并验证了6个机器学习模型,包括CatBoost,使用5倍交叉验证.
- 利用SHAP分析来确定各种因素及其相互作用在毒性预测中的重要性.
主要成果:
- CatBoost 模型实现了高预测性能 (AUC:0.935 ± 0.021,F1:0.832).
- 实验条件 (时间,度),污染物水性和分子基结构显著影响了关节毒性.
- 基和多环芳香化合物被确定为关键的有毒基结构,有助于协同效应.
结论:
- 该研究为MP-有机污染物关节毒性提供了一个强大的预测框架.
- 提出了一个新的协同-对抗性指数来区分联合效应类型.
- 这些发现为水生环境中共同暴露的高通量风险评估提供了洞察力.
更多相关视频
相关概念视频
Green Algae
985
Green algae, also referred to as chlorophytes, are different from red algae in having the chloroplasts containing chlorophylls a and b, which give them their distinct green hue. However, they lack phycobiliproteins, preventing them from developing the red or blue-green pigmentation seen in red algae. In terms of photosynthetic pigment composition, green algae closely resemble plants and share a close evolutionary relationship with them. Taxonomically Green algae belong to Phylum Chlorophyta in...
985
Overview of Algae
1.2K
The kingdom Archaeplastida encompasses red and green algae, along with land plants. Unlike other protists with chloroplasts that arose through secondary endosymbiosis, only red and green algae originated from primary endosymbiotic events. This diverse group of eukaryotic organisms contains chlorophyll and performs oxygenic photosynthesis.Algae exist in various forms, from large brown kelp in coastal waters to green scum in puddles and stains on rocks or soil. Some species are responsible for...
1.2K


