ST ALOYSIUS COLLEGE

Institute of Management and IT- AIMIT

Experimental Design for Microarray

Studies

MSc Bioinformatics Harshit Sajal 231809

JP Sampath Kumar 231810
Stuti Mazgaonkar 231829
Contents

Introduction to DNA Microarray

Microarray and Bioinformatics – Experimental Design

Microarray Data Analysis Using R

DNA
Microarray
Microarray studies are powerful tools for analyzing gene expression across large
numbers of genes.

A DNA microarray (also commonly known as DNA chip or biochip) is a collection

of microscopic DNA spots attached to a solid surface.

Scientists use DNA microarrays to measure the expression levels of large

numbers of genes simultaneously or to genotype multiple regions of a genome.
DNA
Microarray
DNA
Microarray
Microarrays use relative quantitation in which the intensity of a feature is
compared to the intensity of the same feature under a different condition,
and the identity of the feature is known by its position.
DNA
Microarray

Applications of DNA Microarray

• Gene expression profiling

• Comparative genomic hybridization
• Chromatin immunoprecipitation on Chip
• SNP detection
• Alternative splicing detection
• Fusion genes microarray
• Diagnostics and genetic engineering
• Functional genomics
• Toxicological research (Toxicogenomics)
Microarrays and
bioinformatics
The affordability of microarrays led to a surge in their use, but it also presented
several bioinformatics challenges that researchers had to tackle.

1. Multiple Levels of Replication (Experimental Design)

Microarray experiments often involve biological replicates (samples from the same
condition) and technical replicates (repeated measurements of the same sample).
Distinguishing between biological variation and technical noise requires careful design
and analysis.

There are three main elements to consider when designing a microarray experiment.

• Replication of the biological samples is essential for drawing conclusions from the
experiment.

• Technical replicates (e.g. two RNA samples obtained from each experimental unit) may
help to quantitate precision.

• Spots of each cDNA clone or oligonucleotide are present as replicates (at least
duplicates) on the microarray slide, to provide a measure of technical precision in each
hybridization.
Microarrays and
bioinformatics
2. Number of Platforms and Data Formats (Standardization)
Early microarrays were platform-specific, with different manufacturers using varying
probe designs and data formats. This hindered data sharing and comparisons across
studies.

Consortiums like the Microarray Gene Expression Data Society (MGED Society)
established minimum information about a microarray experiment (MIAME) standards.
These guidelines ensure consistent data reporting and facilitate comparisons between
studies.

3. Statistical Treatment of the Data (Data Analysis)

Microarray data is high-dimensional, with thousands of genes measured simultaneously.
Traditional statistical methods weren't well-suited for analyzing such large datasets while
accounting for multiple testing issues.

Statistical challenges include taking into account effects of background noise and
appropriate normalization of the data.
Microarrays and
bioinformatics
Algorithms that affect statistical analysis include: Image analysis, data processing, Class
discovery analysis and prediction, etc.

Identification of statistically significant changes in gene expression are commonly

identified using the t-test, ANOVA, Bayesian method, Mann–Whitney test method and
PCA tailored to microarray data sets, which take into account multiple comparisons.

4. Mapping Probes to mRNA Transcripts (Annotation)

Microarray probes target specific sequences, but linking them to the exact mRNA
transcripts they measure can be challenging due to alternative splicing or probe cross-
hybridization.

5. Data Warehousing and Sharing (Data Sharing)

The sheer volume of microarray data generated presented challenges in storage,
management, and sharing.

Gene Expression Omnibus (GEO) provide a centralized repository for microarray

data in standardized formats.
Microarrays and
bioinformatics
EXPERIMENTAL DESIGNS FOR MICRO ARRAY STUDIES
A. Dye Swap (Balanced Block Design)

• Utilizes two dyes (e.g., Cy3 and Cy5) to label two samples being compared.

• Involves swapping the dyes in a second hybridization to minimize dye-specific biases.

• Helps control for systematic errors and technical variations, enhancing data accuracy.

• Enables the identification of true differential gene expression by accounting for dye
effects.

• Particularly useful for studies aiming at detecting subtle changes in gene expression.

• Facilitates the detection of biological effects by reducing technical artifacts in

microarray data.

• Recommended for experiments where sample labeling and hybridization conditions

can introduce bias.
B. Reference Design

• Includes a common reference sample against which all experimental samples are
compared.

• Allows for direct comparisons between different conditions or treatments.

• Enhances statistical power by reducing variability across arrays and within experiments.

• Useful for identifying genes consistently upregulated or downregulated across multiple

samples.

• Enables the detection of relative changes in gene expression levels under different
conditions.

• Requires careful selection and preparation of the reference sample to ensure data
reliability.

• Suitable for experiments aiming to identify biomarkers or pathways associated with

specific conditions.
C. Simple Loop Design

• Involves a circular or linear arrangement of hybridizations, comparing each

sample with the next.

• Efficient for pairwise comparisons between multiple conditions or time points.

• Reduces the number of hybridizations compared to a full factorial design.

• Suitable for studies with limited resources or focusing on specific comparisons.

• Allows for the identification of condition-specific gene expression patterns.

• Provides insights into differential gene regulation between adjacent samples in

the loop.

• Requires careful planning to ensure balanced comparisons and interpretation of

results.
D. Factorial Design

• Incorporates multiple factors (e.g., treatments, genotypes) and their interactions into
the experimental design.

• Enables the simultaneous study of how different variables influence gene expression.

• Allows for the detection of main effects of factors and their interactions on gene
expression.

• Facilitates the identification of genes responsive to specific combinations of factors.

• Provides a comprehensive view of complex biological responses and regulatory

networks.

• Requires careful experimental design to balance factors and replicate conditions

adequately.

• Recommended for studying multifactorial influences on gene expression and system-

level responses.
E. Time-course Reference Design

• Specifically designed for time-course experiments, where samples are collected at

multiple time points.

• Uses a common reference sample across all time points to normalize gene expression
changes.

• Enables the analysis of temporal patterns in gene expression, such as early response
genes or dynamic regulatory processes.

• Facilitates the identification of genes with coordinated expression changes over time.

• Helps differentiate between immediate and delayed responses to stimuli or treatments.

• Requires precise time-point sampling and synchronization for accurate data

interpretation.

• Suitable for studying biological processes with time-dependent regulatory mechanisms.

F. Direct-mixed Time-course

• Involves direct comparisons between samples from different time points without a
common reference.

• Captures temporal changes in gene expression levels across consecutive or non-

consecutive time intervals.

• Allows for the analysis of gene expression dynamics without potential biases introduced
by a reference sample.

• Enables the identification of time-dependent gene expression patterns or transient

responses.

• Suitable for studying rapid physiological changes, transient signaling events, or short-term
regulatory cascades.

• Requires careful data normalization and statistical methods to account for time-
dependent variations.

• Useful for understanding the kinetics of gene expression changes in response to stimuli or
developmental stages.
Microarrays Data Analysis
using R

We will perform some quality assessment, differential expression and

downstream analysis such as clustering.

GEO GSE33126: Nine paired tumor/normal colon tissues on Illumina

HT12_v3 gene expression Beadchips. Neoplastic samples were obtained
from the central area of the neoplasia.

R Packages BiocManager GEOquery

forcats: limma
ggrepel and ggplot2 Pheatmap
tidyr org.Hs.eg.db
References

• Wikipedia: https://en.wikipedia.org/wiki/Microarray
• DNA Microarray- Definition, Principle, Procedure, Types:
https://microbenotes.com/dna-microarray/
• Wikipedia: https://en.wikipedia.org/wiki/DNA_microarray
• Wilson, Claire H. et al. “Experimental Design and Analysis of Microarray Data.”
Applied Mycology and Biotechnology vol. 6 (2006): 1–36. doi:10.1016/S1874-
5334(06)80004-3
• Analysing data from GEO: https://sbc.shef.ac.uk/geo_tutorial/tutorial.nb.html