Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: May 28, 2026

Mining Spatial Transcriptomics Datasets using DeepSpaceDB
10:16

Mining Spatial Transcriptomics Datasets using DeepSpaceDB

Published on: September 5, 2025

DREAMER-S: Deep leaRning-Enabled Attention-based Multiple-instance approaches with Explainable Representations for

M Rifqi Rafsanjani1,2, Alison Dooney2, Rahul Suresh1,2

  • 1Radiation and Environmental Science Centre, Physical to Life Sciences Research Hub, Technological University Dublin, Dublin, Ireland.

Plos Computational Biology
|May 26, 2026
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Co-clinical CT radiomics pipeline to establish candidate imaging biomarkers for colorectal cancer.

European journal of radiology·2026
Same author

Cell engulfment defines spatially distinct competitive metabolic niches associated with clinical outcomes in colorectal cancer.

Cell death and differentiation·2026
Same author

Spatial analyses implicate high stromal tumour-infiltrating CD8<sup>+</sup> lymphocytes as a negative predictive marker for chemotherapy in estrogen receptor-positive breast cancer.

Nature communications·2026
Same author

Reproducible 3D Glioblastoma Migration Assay with Magnetic Nanoparticle Mediated Spheroid Localization Under Hypoxic Conditions.

Journal of visualized experiments : JoVE·2026
Same author

Beyond motor: a systematic review of multisensory integration deficits in Parkinson's disease.

Journal of neural transmission (Vienna, Austria : 1996)·2026
Same author

Antibody response following rabies vaccination: a retrospective cohort study from a tertiary centre in Kerala, India.

Transactions of the Royal Society of Tropical Medicine and Hygiene·2026
This summary is machine-generated.

DREAMER-S, an attention-based multiple-instance learning framework, identifies key spatial features in 3D imaging for classification without pixel-level data. This method aids spatial biology by highlighting relevant regions in high-content datasets.

Area of Science:

  • Spatial biology
  • Computational pathology
  • Medical imaging analysis

Background:

  • Accurate diagnostic and prognostic classification from large-scale, multi-channel spatial imaging is hindered by the lack of pixel-level annotations.
  • Identifying informative spatial features in 3D imaging hypercubes requires advanced computational methods.

Purpose of the Study:

  • To develop an attention-based multiple-instance learning (MIL) framework, DREAMER-S, for learning informative spatial features from image- or slide-level labels.
  • To demonstrate the framework's capability in highlighting class-relevant spectral instances in Quantum Cascade Laser infrared (QCL-IR) tissue imaging without manual annotation.
  • To establish the broad transferability of DREAMER-S to spatial-biology applications requiring instance-level filtering for salient regions of interest.

Main Methods:

Related Experiment Videos

Last Updated: May 28, 2026

Mining Spatial Transcriptomics Datasets using DeepSpaceDB
10:16

Mining Spatial Transcriptomics Datasets using DeepSpaceDB

Published on: September 5, 2025

  • Developed DREAMER-S, an attention-based MIL framework utilizing 3D imaging hypercubes and image/slide-level labels.
  • Applied DREAMER-S to Quantum Cascade Laser infrared (QCL-IR) tissue imaging, rendering attention weights spatially to identify relevant spectral instances.
  • Evaluated DREAMER-S on a colorectal cancer patient-derived xenograft (PDX) model for chemotherapy response prediction.

Main Results:

  • DREAMER-S successfully identified informative spatial features in QCL-IR imaging, with attention weights highlighting class-relevant spectral instances.
  • The framework achieved an F1 score of ~0.95 in separating chemo-sensitive and less responsive PDX models.
  • Model saliency analysis revealed a mechanistic link between spectral signals and apoptosis pathways, correlating with pro-apoptotic protein measurements.

Conclusions:

  • DREAMER-S provides an efficient and interpretable approach for analyzing high-content spatial-biology imaging datasets.
  • The attention-based MIL framework enables the identification of diagnostically relevant spatial features without requiring pixel-level annotations.
  • The method facilitates a deeper understanding of biological mechanisms by linking spectral signals to cellular physiology and treatment response.