Establishment and implementation of multiplexed error robust fluorescence in situ hybridization to study spatial mRNA expression in cells and tissues

Schrader, Alex William

Permalink

https://hdl.handle.net/2142/122237 Copy

Description

Title

Establishment and implementation of multiplexed error robust fluorescence in situ hybridization to study spatial mRNA expression in cells and tissues

Author(s)

Schrader, Alex William

Issue Date

2023-11-29

Director of Research (if dissertation) or Advisor (if thesis)

Han, Hee-Sun

Doctoral Committee Chair(s)

Han, Hee-Sun

Committee Member(s)

• Zhao, Sihai Dave
• Chan, Jefferson
• Jain, Prashant

Department of Study

Chemistry

Discipline

Chemistry

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Imaging spatial transcriptomics
• Image analysis
• Honey bee brain spatial transcriptomics
• Multiplexed error-robust fluorescence in situ hybridization
• MERFISH

Abstract

Spatial transcriptomics has rapidly grown in the last decade, with the development of several commercial systems to explore the spatial organization of Ribonucleic acid (RNA) molecules in cells and tissues. RNA is a key biomolecule that both encodes protein information and performs regulatory functions in the cell. RNA is an attractive target for measuring the cell state because DNA probes can be easily designed to bind to specific RNA transcripts, unlike protein antibodies which are difficult and expensive to make for an arbitrary target. RNA is often used to identify cell types and measure changes in expression due to biological phenotypes. Multiplexed Error Robust Fluorescence in situ Hybridization (MERFISH) is an imaging spatial transcriptomic method that can visualize RNA expression at subcellular resolution. MERFISH is a sequential FISH experiment, were multiple rounds of staining with fluorescently labeled DNA probes and imaging are used to encode transcript identities of hundreds to thousands of genes in a single experiment. I built an automated fluidics system, including software and hardware, to do this staining and imaging automatically. In addition, I have implemented and adapted the software to detect transcripts from images, and then quantify error statistics. To validate our system, we have imaged U-2 OS cultured cancer cells previously used in the literature, and we have found that our data had very similar error rates and strong correlation to both published MERFISH and bulk RNA-Sequencing data. With our systems and methods validated, we moved on to a new biological system, the honey bee. The honey bee is an interesting biological system due to its complex social structure and wide behavioral repertoire even though it has a very small brain (100,000 times smaller than humans). We and our collaborators in the Robinson lab are interested in how aggression is encoded in the honey bee brain. We have specifically selected genes that are related to aggression in soldier and forager honey bees. We physically expanded the honey bee brains using expansion microscopy because the very small cells present in the honey bee brain leads to transcript densities too high to detect. Using the same image decoding methods as with U-2 OS, we were able to identify millions of transcripts from 130 genes in 8-10 µm thick honey bee brain sections. We confirmed the spatial expression patterns of DopR2 and LHX3 by compar We confirmed the spatial expression patterns of two genes, DopR2 and LHX3, by comparing to existing literature. This confirms that our decoding is accurate. With this confirmed, we then implemented cell segmentation using a machine learning algorithm called Cellpose, and were able to cluster cells into cell types across multiple samples. After batch correction, we found that the spatial locations of clusters were consistent across sections. While exploring the data, we were able to identify neurites and somata in the optic lobe region of the honey bee brain, and identified 33 differentially expressed genes in this area. For example, Nesprin-1 was found to be differentially expressed in the somata of the neuron, and Fibroblast growth factor receptor homolog 1 was found to be differentially expressed in the neurite region of the neuron. To increase the number of genes available for spatial analysis, we have also explored imputation by using single cell expression data collected by the Robinson lab to predict spatial expression patterns. We have focused on better understanding the resulting expression values and exploring validation methods to confirm that the predicted expression is accurate. Here we are able to report the first MERFISH results in the honey bee brain, and more broadly, in the insect brain.

Graduation Semester

2023-12

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/122237

Copyright and License Information

Copyright 2023 Alex Schrader

Owning Collections