Department of Experimental and Clinical Toxicology, Institute of Experimental and Clinical Pharmacology and Toxicology, Saarland University, Homburg (Saar), Germany.
Clin Chem. 2010 Apr;56(4):575-84. doi: 10.1373/clinchem.2009.135517. Epub 2010 Feb 25.
The challenge in systematic toxicological analysis using gas chromatography and/or liquid chromatography coupled to mass spectrometry is to identify compounds of interest from background noise. The large amount of spectral information collected in one full-scan MS run demands the use of automated evaluation of recorded data files. We evaluated the applicability of the freeware deconvolution software AMDIS (Automated Mass Spectral Deconvolution and Identification System) for GC-MS-based systematic toxicological analysis in urine for increasing the speed of evaluation and automating the daily routine workload.
We prepared a set of 111 urine samples for GC-MS analysis by acidic hydrolysis, liquid-liquid extraction, and acetylation. After analysis, the resulting data files were evaluated manually by an experienced toxicologist and automatically using AMDIS with deconvolution and identification settings previously optimized for this type of analysis. The results by manual and AMDIS evaluation were then compared.
The deconvolution settings for the AMDIS evaluation were successfully optimized to obtain the highest possible number of components. Identification settings were evaluated and chosen for a compromise between most identified targets and general number of hits. With the use of these optimized settings, AMDIS-based data analysis was comparable or even superior to manual evaluation and reduced by half the overall analysis time.
AMDIS proved to be a reliable and powerful tool for daily routine and emergency toxicology. Nevertheless, AMDIS can identify only targets present in the user-defined target library and may therefore not indicate unknown compounds that might be relevant in clinical and forensic toxicology.
使用气相色谱和/或液相色谱与质谱联用进行系统毒理学分析的挑战在于从背景噪声中识别出感兴趣的化合物。在一次全扫描 MS 运行中收集的大量光谱信息要求使用自动评估记录的数据文件。我们评估了免费软件去卷积软件 AMDIS（自动质谱解卷积和识别系统）在基于 GC-MS 的尿液系统毒理学分析中的适用性，以提高评估速度并实现日常工作的自动化。
我们通过酸性水解、液液萃取和乙酰化制备了一组 111 个尿液样本进行 GC-MS 分析。分析后，由经验丰富的毒理学家手动评估和自动使用 AMDIS 评估生成的数据文件，同时使用针对此类分析进行了优化的去卷积和识别设置。然后比较手动和 AMDIS 评估的结果。
成功优化了 AMDIS 评估的去卷积设置，以获得尽可能多的成分。评估并选择了识别设置，以在尽可能多的目标和一般命中数量之间取得平衡。使用这些优化的设置，基于 AMDIS 的数据分析与手动评估相当，甚至更优，并将总分析时间缩短了一半。
AMDIS 被证明是日常和紧急毒理学的可靠且强大的工具。然而，AMDIS 只能识别用户定义的目标库中存在的目标，因此可能无法指示在临床和法医毒理学中可能相关的未知化合物。
