In Situ Hybridization (ISH) and NC2 Analysis

• NC2 microscopic analysis advantages: non-destructive visualization, real-time monitoring
  • Sampling speed: millisecond timescale with modern microscopic cameras
  • Sensitivity: single molecule fluorescence detection possible
  • Resolution: typically 1 micron or 0.25 micron (diffraction limited)
  • Can achieve 10 nanometers using near field optics tricks
• Major limitation: multiplicity (1-3 colors typically)
  • Can discriminate 24 human metaphase chromosomes using color combinations
  • Practical limit: 4-5 colors due to overlap issues
  • Not immediately genomics-compatible for systems biology

Microarray Technologies Overview

• Long probes (microarrays): entire gene/cDNA length
  • Single probe per gene
  • Two-color experiments (experiment vs control)
  • Mechanically spotted
• Short probes (oligonucleotides): 25 nucleotides
  • 20 probes per gene
  • Photochemical synthesis using photolithography
  • Perfect match (PM) and mismatch (MM) controls
  • Subtract MM from PM, average across 20 probes
• Quantitation methods:
  • Top methods: fluorescent signal integration
  • Bottom methods: individual tag counting

Sequencing-Based Methods

• SAGE (Serial Analysis of Gene Expression)
• MPSS (Massively Parallel Signature Sequencing)
• Both sequence 14-22 nucleotides
  • 14-mer: sufficient for cDNA database recognition
  • 20-mer: needed for genomic database accuracy
• Advantage: discovery capability vs predetermined array features

Microarray Technical Details

• Photolithography arrays: ~1 million features per square centimeter
• Each feature: 10^5 to 10^6 identical molecules
• Sample preparation: biotinylation or direct fluorescent labeling
• Kinetic binding based on mass action principles
• Quantitation via streptavidin-fluorophore conjugates

Experimental Reproducibility and Error Analysis

• Early experiments assumed single-use arrays (cost-driven but not effective)
• Reproducibility assessment crucial for reliable results
• High copy number RNAs: very close similarity between replicates
• Low copy number RNAs: noise dominates (light scattering, background fluorescence, cross-hybridization)
• Yeast data: large fraction of RNAs present at single copy per cell
• Biological significance vs technical noise considerations

Statistical Analysis Methods

• Student t-test for comparing distributions
  • t-statistic = mean difference / standard deviation
  • Null hypothesis: no difference between distributions
  • Requires normal distribution assumption
• Non-parametric alternatives: Wilcoxon matched pairs sign ranks test
  • Lower power but fewer assumptions
  • Used when normality questionable
• Multiple hypothesis testing correction required
  • Significance threshold must account for number of tests
  • Need sharp biological questions or improved data quality

Microarray Data Quality Assessment

• Independent experiments definition critical
  • True independence: separate cell cultures, different researchers
  • Technical replicates: same RNA sample, different labeling
• R-squared values decrease with increasing experimental independence
• Guidelines for RNA quantitation:
  • Early limit: 3-fold effects
  • Biological motivation: 1.5-fold effects (human trisomy examples)
  • Goal: detect biologically significant 1.5-fold changes

Oligonucleotide Advantages and Considerations

• Higher multiplicity per gene possible
• Specificity for individual splice forms
• Cannot determine cis relationships between exons
• Economic pressure toward genome subsets limits meta-analysis capabilities
• Coefficient of variation analysis:
  • >20% CV: less trustworthy
  • Reproducibility enables detection of low-abundance RNAs
  • Multiple experiments beat single-experiment software calls

Alternative Quantitation Methods Comparison

• Immobilized genes/labeled RNA (microarrays): high throughput, high multiplicity
• ISH: spatial relationships preserved
• Immobilized RNA/labeled probes: size information available
• RT-qPCR: highest sensitivity (10^-4 copies per cell detection)
• Reporter constructs: no cross-hybridization issues
• Counting methods: gene discovery capability

Sources of Systematic and Random Errors

• Secondary structure effects on hybridization efficiency
• Spatial position effects on arrays
• Variable target nucleic acid immobilization
• Cross-hybridization from related sequences
• Unanticipated transcripts in "non-coding" regions
• Solution: internal standards, statistical modeling, spatial randomization

Advanced Array Applications

• Tiling arrays for gene discovery
• Secondary structure mapping using genomic DNA controls
• Transcriptional start site mapping
• Splice junction analysis and exon boundary refinement
• Example: Chromosome 22 tiling study
  • 60-mer oligonucleotides every 10 base pairs
  • Identified annotation errors in exon boundaries
  • Confirmed splice site motifs (GT donor, AG acceptor)

Time Series Analysis

• Advantages over knockout studies: primary vs secondary effects
• Conditional control methods:
  • Chemical knockouts (e.g., rifampicin for RNA polymerase)
  • Temperature-sensitive mutants
  • Avoid broad perturbations
• RNA stability measurements: half-lives from 2 minutes to multiple cell generations
• Dynamic time warping for aligning different time series
  • Similar to sequence alignment dynamic programming
  • Accounts for different time constants and warping
  • Cell cycle synchronization example using different methods

Cell Cycle Synchronization Applications

• Addresses cell population heterogeneity
• Multiple synchronization methods create alignment challenges
• Dynamic programming solutions for time series alignment
• Rich datasets from thousands of genes enable pattern recognition