Colocalization analysis for multimodal optical microscopy

Novel microscopy techniques allow for the acquisition of multiple modalities of biological images simultaneously. Label-free optical methods take advantage of the molecular, structural, and chemical makeup of biological samples to collect endogenous contrasts without the need for harmful or destructive dyes. While multimodal label-free optical imaging enables many new kinds of biological experiments, there exists an analysis bottleneck between data collection and biological discovery. By calculating “colocalization” metrics to quantify relationships between complementary spatially and temporally co-registered channels, these rich relationships that exist across modalities can begin to be understood. First, synthetic data was used to motivate this exploration, showing that there does not exist one metric that can capture all the various kinds of multimodal relationships that are common in biological data. Then, the significance and value of colocalization features in a breast cancer dataset was explored and its effectiveness against handcrafted morphological features was benchmarked. It was found that not only are there relationships across features and modalities, but by combining information across channels with that within each channel, spatially varying colocalization patterns can be observed. Synthesizing these findings, a colocalization analysis pipeline was applied to three datasets at the tissue-, cellular-, and subcellular-levels to analyze the complex relationships held across channels. Through a thorough study of colocalization signals, it was seen that that inter-channel colocalization relationships contain just as much information as intra-channel morphological relationships in biological tasks, and analysis methods must be carefully tailored to the biological task of interest.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Rao, Yug

Contributors dc:contributor

• Boppart, Stephen A.

• Colocalization
• Microscopy
• Label-free Imaging
• Image Analysis

Statement dc:rights

• Copyright 2025 Yug Rao

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/130067

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F130067