Targeted proteomics data interpretation with DeepMRM

Area of Science:

• Biochemistry
• Computational Biology
• Proteomics

Background:

• Targeted proteomics is crucial in clinical research but suffers from time-consuming manual data interpretation.
• This manual process limits the transferability, reproducibility, and scalability of targeted proteomics approaches.
• Automating data analysis is essential for advancing clinical proteomics applications.

Purpose of the Study:

• To introduce DeepMRM, a novel deep learning-based software package for automated targeted proteomics data analysis.
• To reduce manual intervention in the interpretation of targeted proteomics data.
• To improve the accuracy, reproducibility, and scalability of targeted proteomics.

Main Methods:

• Developed DeepMRM using deep learning algorithms for object detection.
• Evaluated DeepMRM performance on both internal and public targeted proteomics datasets.
• Compared DeepMRM's accuracy against the established tool Skyline.

Main Results:

• DeepMRM demonstrated superior accuracy in targeted proteomics data analysis compared to Skyline.
• The software effectively minimizes the need for manual data interpretation.
• DeepMRM offers enhanced accuracy and efficiency for analyzing complex proteomics data.

Conclusions:

• DeepMRM significantly improves the automation and accuracy of targeted proteomics data analysis.
• The software addresses key limitations in reproducibility and scalability faced by current methods.
• DeepMRM is available as a stand-alone tool and integrated into Skyline to encourage adoption.