Self-supervised clustering of mass spectrometry imaging data using contrastive learning

Hu, Hang; Bindu, Jyothsna Padmakumar; Laskin, Julia

doi:10.1039/d1sc04077d

Cited by 16 publications

(15 citation statements)

References 37 publications

(44 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multiple ion channels are used in the DLADS training, testing, and implementation to obtain accurate molecular distributions for different types of molecules observed experimentally. Specifically, several representative m / z images with distinct molecular distributions can be selected using a self-supervised molecular colocalization clustering approach . Pixels with the highest mean ERD values from all m/z channels are then preferentially sampled in acquisition steps.…”

Section: Theoretical Methodsmentioning

confidence: 99%

“…Specifically, several representative m/z images with distinct molecular distributions can be selected using a self-supervised molecular colocalization clustering approach. 45 Pixels with the highest mean ERD values from all m/z channels are then preferentially sampled in acquisition steps. For the simulation study, the DLADS model was trained with RD formed independently for each m/z channel, where during testing and implementation, the ERD matrices of each m/z channel were averaged together using eq 4.…”

Section: ■ Theoretical Methodsmentioning

confidence: 99%

See 1 more Smart Citation

High-Throughput Mass Spectrometry Imaging with Dynamic Sparse Sampling

Helminiak

Yang

et al. 2022

ACS Meas. Sci. Au

Self Cite

View full text Add to dashboard Cite

Mass spectrometry imaging (MSI) enables label-free mapping of hundreds of molecules in biological samples with high sensitivity and unprecedented specificity. Conventional MSI experiments are relatively slow, limiting their utility for applications requiring rapid data acquisition, such as intraoperative tissue analysis or 3D imaging. Recent advances in MSI technology focus on improving the spatial resolution and molecular coverage, further increasing the acquisition time. Herein, a deep learning approach for dynamic sampling (DLADS) was employed to reduce the number of required measurements, thereby improving the throughput of MSI experiments in comparison with conventional methods. DLADS trains a deep learning model to dynamically predict molecularly informative tissue locations for active mass spectra sampling and reconstructs high-fidelity molecular images using only the sparsely sampled information. Experimental hardware and software integration of DLADS with nanospray desorption electrospray ionization (nano-DESI) MSI is reported for the first time, which demonstrates a 2.3-fold improvement in throughput for a linewise acquisition mode. Meanwhile, simulations indicate that a 5–10-fold throughput improvement may be achieved using the pointwise acquisition mode.

show abstract

Section: Theoretical Methodsmentioning

confidence: 99%

Section: ■ Theoretical Methodsmentioning

confidence: 99%

High-Throughput Mass Spectrometry Imaging with Dynamic Sparse Sampling

Helminiak

Yang

et al. 2022

ACS Meas. Sci. Au

Self Cite

View full text Add to dashboard Cite

show abstract

“…Adapted with permission. [158] Copyright 2022, Royal Society of Chemistry. though many machine learning algorithms had been tested, including PCA, LDA, SVM, and RF, the recurring image recognition challenge has been addressed only recently using deep CNN.…”

Section: Data Compression For Spatial Datamentioning

confidence: 99%

“…Self-supervised learning, which does not require any annotated data, is a promising approach for addressing this challenge. Hu et al [158] have developed a selfsupervised clustering approach for MSI spatial data compression using contrastive learning and image augmentation. Based on a simple framework for contrastive learning of visual representations (SimCLR) shown in Figure 7A, a deep CNN encoder was trained using mouse uterine MSI data from a single experiment without manual annotations.…”

Section: Data Compression For Spatial Datamentioning

confidence: 99%

“…B) A t-SNE plot of learned image representations, in which each data point corresponds to an ion image and is color-coded by the manually annotated spatial pattern classifications used as ground truth (0-12).Representative ion images are also shown around the scatter plot and linked to their locations on the t-SNE map. Adapted with permission [158]. Copyright 2022, Royal Society of Chemistry.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Emerging Computational Methods in Mass Spectrometry Imaging

Laskin

2022

Advanced Science

Self Cite

View full text Add to dashboard Cite

Mass spectrometry imaging (MSI) is a powerful analytical technique that generates maps of hundreds of molecules in biological samples with high sensitivity and molecular specificity. Advanced MSI platforms with capability of high-spatial resolution and high-throughput acquisition generate vast amount of data, which necessitates the development of computational tools for MSI data analysis. In addition, computation-driven MSI experiments have recently emerged as enabling technologies for further improving the MSI capabilities with little or no hardware modification. This review provides a critical summary of computational methods and resources developed for MSI data analysis and interpretation along with computational approaches for improving throughput and molecular coverage in MSI experiments. This review is focused on the recently developed artificial intelligence methods and provides an outlook for a future paradigm shift in MSI with transformative computational methods.

show abstract

MatGPT: A Vane of Materials Informatics from Past, Present, to Future

Wang,

Chen,

Tao

et al. 2023

Advanced Materials

View full text Add to dashboard Cite

Combining materials science, artificial intelligence (AI), physical chemistry and other disciplines, materials informatics is continuously accelerating the vigorous development of new materials. The emergence of “GPT AI” shows that the scientific research field has entered the era of intelligent civilization with “data” as the basic factor and “algorithm + computing power” as the core productivity. The continuous innovation of AI will impact the cognitive laws and scientific methods, and reconstruct the knowledge and wisdom system. This leads us to think more about materials informatics, both in terms of opportunities and challenges. In this review, we provide a comprehensive discussion of AI models, materials infrastructures, and review the current advances in the discovery and design of new materials. With the rise of new research paradigms triggered by “AI for Science”, we propose the vane of materials informatics: “MatGPT”, and carry out the technical path planning from the aspects of data, descriptors, generative models, pre‐training models, directed design models, collaborative training, experimental robots, as well as the efforts and preparations needed to develop a new generation of materials informatics. Finally, we discuss the challenges and constraints faced by materials informatics, in order to achieve a more digital, intelligent and automated construction of materials informatics with the joint efforts of more interdisciplinary scientists. This article is protected by copyright. All rights reserved

show abstract

Self-supervised clustering of mass spectrometry imaging data using contrastive learning

Abstract: Contrastive learning is used to train a deep convolutional neural network to identify high-level features in mass spectrometry imaging data. These features enable self-supervised clustering of ion images without manual annotation.

Cited by 16 publications

References 37 publications

High-Throughput Mass Spectrometry Imaging with Dynamic Sparse Sampling

High-Throughput Mass Spectrometry Imaging with Dynamic Sparse Sampling

Emerging Computational Methods in Mass Spectrometry Imaging

MatGPT: A Vane of Materials Informatics from Past, Present, to Future

Contact Info

Product

Resources

About