Colocalization Features for Classification of Tumors Using Desorption Electrospray Ionization Mass Spectrometry Imaging.

Paolo Inglese,Robert C Glen,Gonçalo Correia,Zoltan Takats,Pamela Pruski

doi:10.1021/acs.analchem.8b05598

Paolo Inglese, Robert C Glen + Show 3 more

Open Access

https://doi.org/10.1021/acs.analchem.8b05598

Copy DOI

Journal: Analytical Chemistry	Publication Date: Apr 23, 2019
Citations: 17	License type: cc-by

Affiliation: Imperial College London

Abstract

Supervised modeling of mass spectrometry imaging (MSI) data is a crucial component for the detection of the distinct molecular characteristics of cancerous tissues. Currently, two types of supervised analyses are mainly used on MSI data: pixel-wise segmentation of sample images and whole-sample-based classification. A large number of mass spectra associated with each MSI sample can represent a challenge for designing models that simultaneously preserve the overall molecular content while capturing valuable information contained in the MSI data. Furthermore, intensity-related batch effects can introduce biases in the statistical models. Here we introduce a method based on ion colocalization features that allows the classification of whole tissue specimens using MSI data, which naturally preserves the spatial information associated the with the mass spectra and is less sensitive to possible batch effects. Finally, we propose data visualization strategies for the inspection of the derived networks, which can be used to assess whether the correlation differences are related to coexpression/suppression or disjoint spatial localization patterns and can suggest hypotheses based on the underlying mechanisms associated with the different classes of analyzed samples.

Highlights

Supervised modeling of mass spectrometry imaging (MSI) data is a crucial component for the detection of the distinct molecular characteristics of cancerous tissues
Applying any of these techniques to a sample generates a set of ion images that we denote as mass spectrometry (MS) images or, equivalently, MSI sample in this text
Unsupervised methods try to capture the intrinsic statistical properties of the MSI data, such as spectral similarity, and generate partitions of the data without relying on any external ground truth. These methods can be useful to suggest novel hypotheses on the metabolic pathways associated with cancerous tissue, for instance, when histological or other expert-driven properties of the analyzed samples are missing.[15−20] In contrast, supervised methods aim to determine statistical relationships between the observed ion signatures and “labels” associated with the analyzed data, that are manually generated by experts

Summary

Introduction

Supervised modeling of mass spectrometry imaging (MSI) data is a crucial component for the detection of the distinct molecular characteristics of cancerous tissues. Unsupervised methods try to capture the intrinsic statistical properties of the MSI data, such as spectral similarity, and generate partitions of the data without relying on any external ground truth These methods can be useful to suggest novel hypotheses on the metabolic pathways associated with cancerous tissue, for instance, when histological or other expert-driven properties of the analyzed samples are missing.[15−20] In contrast, supervised methods aim to determine statistical relationships between the observed ion signatures and “labels” associated with the analyzed data, that are manually generated by experts. Another approach is based on selecting a random subset of pixel spectra from the tissue of interest and predicting their class.[27]

Methods

Results

Conclusion