Self-supervised Clustering of Mass Spectrometry Imaging Data Using Contrastive Learning

Abstract

Mass spectrometry imaging (MSI) is widely used for the label-free molecular mapping of biological samples. The identification of co-localized molecules in MSI data is crucial to the understanding of biochemical pathways. However, complex MSI data are too large for manual annotation but too small for training deep networks. Herein, we introduce a self-supervised clustering approach based on contrastive learning, which shows an excellent performance in clustering of small MSI data. We train a deep convolutional neural network (CNN) using MSI data from a single experiment without manual annotations to effectively learn high-level spatial features from ion images and classify them based on molecular colocalizations. We demonstrate that contrastive learning generates ion image representations that form well-resolved clusters. Subsequent self-labeling is used to fine-tune both the CNN encoder and linear classifier based on confidently classified ion images. This new approach enables autonomous and highthroughput identification of co-localized species in MSI data, which will dramatically expand the application of spatial lipidomics, metabolomics, and proteomics in biological research.

High-Throughput Mass Spectrometry Imaging with Dynamic Sparse Sampling

“…Specifically, several representative m/z images with distinct molecular distributions can be selected using a self-supervised molecular colocalization clustering approach. 41 During training, the RD and ERD values for each m/z channel were computed to optimize 𝑤 . For DLADS testing and implementation, the ERD matrices of each m/z channel are averaged together using Eq (4).…”

MDDI-SCL: predicting multi-type drug-drug interactions via supervised contrastive learning

“…Contrastive learning has been successfully applied in the field of bioinformatics [33][34][35][36][37][38]. In this study, we propose a new method named MDDI-SCL for multi-type DDI prediction, which is based on Supervised Contrastive Learning (SCL) and three-level loss functions.…”

Molecular Imaging of Biological Samples in Pharmaceutical Development Using Mass Spectrometry Imaging and Machine Learning

“…In addition, a self-supervised clustering approach has been developed to autonomously classify molecules with respect to their spatial distributions using convolutional neural networks. Identification of these functionally related molecules helps analyze biochemical pathways in the tissue [4]. We will demonstrate their applications to study glucuronidation of drugs in mouse kidney [5].…”