Illumina Sequencing

Illumina is the dominant next-generation sequencing platform worldwide. It uses reversible terminator chemistry and fluorescent detection to sequence millions of DNA fragments simultaneously with high accuracy.

The Chemistry: Reversible Terminators

The Core Principle

Unlike Ion Torrent (which detects H⁺ ions), Illumina detects fluorescent light from labeled nucleotides.

Key innovation: Reversible terminators

Normal dNTP:

• Has 3'-OH group
• Polymerase adds it and continues to next base

Reversible terminator (Illumina):

• Has 3'-OH blocked by a chemical group
• Has fluorescent dye attached
• Polymerase adds it and stops
• After imaging, the block and dye are removed
• Polymerase continues to next base

Why this matters: You get exactly one base added per cycle, making base calling precise.

How It Works: Step by Step

1. Library Preparation

DNA is fragmented and adapters are ligated to both ends of each fragment.

Adapters contain:

• Primer binding sites
• Index sequences (barcodes for sample identification)
• Sequences complementary to flow cell oligos

2. Cluster Generation (Bridge Amplification)

This is Illumina's signature step - amplification happens on the flow cell surface.

The flow cell:

• Glass slide with millions of oligos attached to the surface
• Two types of oligos (P5 and P7) arranged in a lawn

Bridge amplification process:

Step 1: DNA fragments bind to flow cell oligos (one end attaches)

Step 2: The free end bends over and binds to nearby oligo (forms a "bridge")

Step 3: Polymerase copies the fragment, creating double-stranded bridge

Step 4: Bridge is denatured (separated into two strands)

Step 5: Both strands bind to nearby oligos and repeat

Result: Each original fragment creates a cluster of ~1,000 identical copies in a tiny spot on the flow cell.

Why amplification? Like Ion Torrent, a single molecule's fluorescent signal is too weak to detect. A thousand identical molecules in the same spot produce a strong signal.

Visual representation:

Original fragment: ═══DNA═══

After bridge amplification:
║ ║ ║ ║ ║ ║ ║ ║
║ ║ ║ ║ ║ ║ ║ ║  ← ~1000 copies in one cluster
║ ║ ║ ║ ║ ║ ║ ║
Flow cell surface

3. Sequencing by Synthesis

Now the actual sequencing begins.

Cycle 1:

1. Add fluorescent reversible terminators (all four: A, C, G, T, each with different color)
2. Polymerase incorporates one base (only one because it's a terminator)
3. Wash away unincorporated nucleotides
4. Image the flow cell with laser
   • Green light = A was added
   • Blue light = C was added
   • Yellow light = G was added
   • Red light = T was added
5. Cleave off the fluorescent dye and the 3' blocking group
6. Repeat for next base

Cycle 2, 3, 4... 300+: Same process, one base at a time.

Key difference from Ion Torrent:

• Illumina: All four nucleotides present at once, polymerase chooses correct one
• Ion Torrent: One nucleotide type at a time, polymerase adds it only if it matches

Color System

2 color and 4 colors system

No Homopolymer Problem

Why Illumina Handles Homopolymers Better

Remember Ion Torrent's main weakness? Homopolymers like AAAA produce strong signals that are hard to quantify (is it 3 A's or 4?).

Illumina doesn't have this problem because:

1. One base per cycle - the terminator ensures only one nucleotide is added
2. Direct counting - if you see 4 green signals in a row, it's exactly 4 A's
3. No signal intensity interpretation - just presence/absence of color

Example:

Sequence: AAAA

Illumina:

Cycle 1: Green (A)
Cycle 2: Green (A)
Cycle 3: Green (A)
Cycle 4: Green (A)
→ Exactly 4 A's, no ambiguity

Ion Torrent:

Flow A: Large signal (proportional to 4 H⁺ ions)
→ Is it 4? Or 3? Or 5? (requires signal quantification)

Error Profile: Substitutions, Not Indels

Illumina's Main Error Type

Substitution errors - reading the wrong base (A instead of G, C instead of T)

Error rate: ~0.1% (1 error per 1,000 bases, or 99.9% accuracy)

Common causes:

1. Phasing/pre-phasing - some molecules in a cluster get out of sync
2. Dye crosstalk - fluorescent signals bleed between channels
3. Quality degradation - accuracy decreases toward end of reads

Why Few Indels?

Because of the reversible terminator:

• Exactly one base per cycle
• Can't skip a base (would need terminator removal without incorporation)
• Can't add two bases (terminator blocks second addition)

Comparison:

Phasing and Pre-phasing

The Synchronization Problem

In a perfect world, all molecules in a cluster stay perfectly synchronized - all at the same base position.

Reality: Some molecules lag behind (phasing) or jump ahead (pre-phasing).

Phasing (Lagging Behind)

Cycle 1: All molecules at position 1 ✓
Cycle 2: 98% at position 2, 2% still at position 1 (incomplete extension)
Cycle 3: 96% at position 3, 4% behind...

As cycles progress, the cluster becomes a mix of molecules at different positions.

Result: Blurry signal - you're imaging multiple bases at once.

Pre-phasing (Jumping Ahead)

Cause: Incomplete removal of terminator or dye

A molecule might:

• Have terminator removed
• BUT dye not fully removed
• Next cycle adds another base (now 2 bases ahead of schedule)

Impact on Quality

Early cycles (1-100): High accuracy, minimal phasing
Middle cycles (100-200): Good accuracy, some phasing
Late cycles (200-300+): Lower accuracy, significant phasing

Quality scores decline with read length. This is why:

• Read 1 (first 150 bases) typically has higher quality than Read 2
• Paired-end reads are used (sequence both ends, higher quality at each end)

Paired-End Sequencing

What Is Paired-End?

Instead of sequencing only one direction, sequence both ends of the DNA fragment.

Process:

1. Read 1: Sequence from one end (forward direction) for 150 bases
2. Regenerate clusters (bridge amplification again)
3. Read 2: Sequence from the other end (reverse direction) for 150 bases

Result: Two reads from the same fragment, separated by a known distance.

Why Paired-End?

1. Better mapping

• If one end maps ambiguously, the other might be unique
• Correct orientation and distance constrain mapping

2. Detect structural variants

• Deletions: Reads closer than expected
• Insertions: Reads farther than expected
• Inversions: Wrong orientation
• Translocations: Reads on different chromosomes

3. Improve assembly

• Links across repetitive regions
• Spans gaps

4. Quality assurance

• If paired reads don't map correctly, flag as problematic

Illumina Systems

Different Throughput Options

Illumina offers multiple sequencing platforms for different scales:

Key trade-offs:

• Higher throughput = longer run time
• Longer reads = lower throughput or longer run time
• Bigger machines = higher capital cost but lower cost per Gb

Advantages of Illumina

1. High Accuracy

• 99.9% base accuracy (Q30 or higher)
• Few indel errors
• Reliable base calling

2. High Throughput

• Billions of reads per run
• Suitable for whole genomes at population scale

3. Low Cost (at scale)

• ~$5-10 per Gb for high-throughput systems
• Cheapest for large projects

4. Mature Technology

• Well-established protocols
• Extensive bioinformatics tools
• Large user community

5. Flexible Read Lengths

• 50 bp to 300 bp
• Single-end or paired-end

6. Multiplexing

• Sequence 96+ samples in one run using barcodes
• Reduces cost per sample

Disadvantages of Illumina

1. Short Reads

• Maximum ~300 bp (vs. PacBio: 10-20 kb)
• Hard to resolve complex repeats
• Difficult for de novo assembly of large genomes

2. Run Time

• 12-44 hours for high-throughput systems
• Longer than Ion Torrent (2-4 hours)
• Not ideal for ultra-rapid diagnostics

3. PCR Amplification Bias

• Bridge amplification favors certain sequences
• GC-rich or AT-rich regions may be underrepresented
• Some sequences difficult to amplify

4. Equipment Cost

• NovaSeq: $850,000-$1,000,000
• High upfront investment
• Requires dedicated space and trained staff

5. Phasing Issues

• Quality degrades with read length
• Limits maximum usable read length

When to Use Illumina

Ideal Applications

Whole Genome Sequencing (WGS)

• Human, animal, plant genomes
• Resequencing (alignment to reference)
• Population genomics

Whole Exome Sequencing (WES)

• Capture and sequence only coding regions
• Clinical diagnostics
• Disease gene discovery

RNA Sequencing (RNA-seq)

• Gene expression profiling
• Transcript discovery
• Differential expression analysis

ChIP-Seq / ATAC-Seq

• Protein-DNA interactions
• Chromatin accessibility
• Epigenomics

Metagenomics

• Microbial community profiling
• 16S rRNA sequencing
• Shotgun metagenomics

Targeted Panels

• Cancer hotspot panels
• Carrier screening
• Pharmacogenomics

Not Ideal For

Long-range phasing (use PacBio or Oxford Nanopore)
Structural variant detection (short reads struggle with large rearrangements)
Ultra-rapid turnaround (use Ion Torrent for speed)
De novo assembly of repeat-rich genomes (long reads better)

Illumina vs Ion Torrent: Summary

The Bottom Line

Illumina is the workhorse of genomics. It's not the fastest (Ion Torrent), not the longest reads (PacBio/Nanopore), but it hits the sweet spot of:

• High accuracy
• High throughput
• Reasonable cost
• Mature ecosystem

For most genomic applications - especially resequencing, RNA-seq, and exomes - Illumina is the default choice.

The main limitation is short reads. For applications requiring long-range information (phasing variants, resolving repeats, de novo assembly), you'd combine Illumina with long-read technologies or use long-read platforms alone.

Key takeaway: Illumina's reversible terminator chemistry elegantly solves the homopolymer problem by ensuring exactly one base per cycle, trading speed (longer run time) for accuracy (99.9%).

About Course Materials

These notes contain NO copied course materials. Everything here is my personal understanding and recitation of concepts, synthesized from publicly available resources (Illumina documentation, sequencing technology literature, bioinformatics tutorials).

This is my academic work—how I've processed and reorganized information from legitimate sources. I take full responsibility for any errors in my understanding.

If you believe any content violates copyright, contact me at mahmoudahmedxyz@gmail.com and I'll remove it immediately.