Whole Genome Sequencing — Day 2: Genome Assembly, Quality Assessment, and Topology

Created in January 08, 2026

2026

🧬 Whole Genome Sequencing — Day 2

Genome Assembly, Quality Assessment, and Topology

In Day 1, we focused on raw read quality control and preprocessing. Today, we move to the next critical step in whole genome analysis:

👉 Genome assembly and assembly evaluation

This step transforms high-quality sequencing reads into contiguous genomic sequences that form the foundation for taxonomy, annotation, and comparative genomics.

🧩 From Reads to Genomes: What Are Reads, Contigs, and Sequences?

Before diving into assembly tools, it’s important to clarify a few core concepts that often cause confusion.

🔹 What are raw reads?

Reads are short (Illumina) or long (Nanopore/PacBio) DNA fragments directly produced by sequencing machines. They represent small, unordered pieces of the genome — like shattered fragments of a book.

• Illumina: typically 100–300 bp

• Nanopore/PacBio: thousands to millions of bp

*On their own, reads do not represent genes or genomes.

🔹 What is a contig?

A contig (contiguous sequence) is created by assembling overlapping reads into a longer continuous DNA sequence.

• Built computationally using assemblers (SPAdes, MEGAHIT, Flye, etc.)

• Represents a reconstructed region of the genome

• Can range from a few kb to several Mb

👉 In draft assemblies, a genome may consist of many contigs.

🔹 What does “sequence” mean in genome assembly?

In practice, a sequence usually refers to:

• a contig, or

• a scaffold, or

• a complete chromosome (single contig)

So when we say “one sequence represents the chromosome”, we mean: ✔ the entire chromosome is assembled into one contig, often circular in bacteria.

##🧠 Assembling a Broken Book

Think of genome assembly like this:

• 📄 Reads → torn sentences

• 📚 Contigs → reconstructed paragraphs

• 📖 Genome → the full book

*Short-read data often leaves gaps (many contigs), while long reads can reconstruct the whole book in one piece. Genome assembly

##🔧 Assembly Strategies Overview

Different sequencing technologies require different assembly approaches:

Data type Typical tools
Illumina short reads SPAdes, MEGAHIT, Shovill
Nanopore / PacBio long reads Flye
Hybrid (short + long) SPAdes hybrid, Unicycler

🧱 Short-Read Genome Assembly (Illumina)

1️⃣ SPAdes

SPAdes is widely used for bacterial isolate assemblies and supports multiple modes.

spades.py \
  -1 sample_R1.fastq.gz \
  -2 sample_R2.fastq.gz \
  -o spades_out \
  --threads 32 \
  --memory 128

Best for: ✔ Isolates ✔ Moderate genome sizes ✔ High accuracy

2️⃣ MEGAHIT

MEGAHIT is optimized for speed and large datasets.

megahit \
  -1 sample_R1.fastq.gz \
  -2 sample_R2.fastq.gz \
  -o megahit_out \
  --num-cpu-threads 32

Best for: ✔ Large datasets ✔ Rapid exploratory assemblies ✔ Metagenomes or many isolates

3️⃣ Shovill (Pipeline wrapper)

Shovill is a fast, practical wrapper around SPAdes with automatic preprocessing.

shovill \
  --R1 sample_R1.fastq.gz \
  --R2 sample_R2.fastq.gz \
  --outdir shovill_out \
  --cpus 32

For more details about Shovil Assembler ; Please read https://jojyjohn28.github.io/blog/genome-assembly-day/ Why I like Shovill: ✔ Simple ✔ Opinionated defaults ✔ Ideal for batch isolate assembly

🧬 Long-Read Assembly (Nanopore / PacBio)

4️⃣ Flye

Flye is designed specifically for long noisy reads and can resolve complete circular chromosomes.

flye \
  --nano-raw sample_nanopore.fastq.gz \
  --out-dir flye_out \
  --threads 32

Flye often outputs:

✔ Single contig bacterial chromosomes

✔ Assembly graphs indicating circular topology

✔ Coverage estimates

For more details; Please read https://jojyjohn28.github.io/blog/genome-assembly-day/

5️⃣ EPI2ME: Real-Time Analysis for Oxford Nanopore Data

EPI2ME is Oxford Nanopore Technologies’ cloud-based (and increasingly local) analysis platform designed to process Nanopore sequencing data in real time.

Why EPI2ME is useful

•Runs during sequencing (live basecalling + analysis)

•Minimal setup — ideal for quick checks and teaching

Provides:

•Read quality summaries

•Taxonomic classification

•Assembly workflows (selected pipelines)

•AMR and functional screens (workflow-dependent)

Limitations

•Less customizable than command-line workflows

•Not ideal for large comparative projects

•Cloud dependency (unless using EPI2ME Labs locally)

Best practice: Treat EPI2ME as a first look, not a replacement for reproducible, command-line assembly pipelines (Flye, SPAdes, etc.).

🧬 Hybrid Assembly: Combining Short and Long Reads

Modern bacterial genome assembly often benefits from hybrid assembly, which integrates:

•Short reads (Illumina) → high base accuracy

•Long reads (ONT / PacBio) → long-range continuity

Together, they solve problems that each technology alone cannot.

Why hybrid assembly matters | Challenge | Short reads | Long reads | Hybrid | | ——————— | ———– | ———- | —— | | Repeats | ❌ | ✅ | ✅ | | Structural resolution | ❌ | ✅ | ✅ | | Base accuracy | ✅ | ❌ | ✅ | | Circular genomes | ❌ | ✅ | ✅ |

🔧 Common Hybrid Assembly Strategies

1️⃣ Short-read–first (polish later)

•Assemble long reads with Flye

•Polish using Illumina reads (e.g., Pilon)

•Best when long-read coverage is high

2️⃣ True hybrid assemblers

• SPAdes (–hybrid)

• Unicycler (highly recommended for bacteria)_ my favorite

Example (Unicycler):

unicycler \
  -1 reads_R1.fastq.gz \
  -2 reads_R2.fastq.gz \
  -l nanopore.fastq.gz \
  -o unicycler_out \
  -t 32

🧠 Choosing the Right Strategy | Data available | Recommended approach | | ———————— | ————————– | | Illumina only | SPAdes / Shovill | | ONT only (high coverage) | Flye | | Illumina + ONT | Unicycler or SPAdes hybrid | | Training / rapid checks | EPI2ME |

Hybrid Genome assembly In this image Left panel: fragmented / problematic assembly and Right panel: improved assembly after hybrid approach. The assembly graph is visulized on Bandage.

📊 Assembly Evaluation with QUAST

Once assemblies are generated, QUAST is used to evaluate their quality.

quast.py \
  spades_out/contigs.fasta \
  megahit_out/final.contigs.fa \
  shovill_out/contigs.fa \
  flye_out/assembly.fasta \
  -o quast_results

QUAST reports include:

• Genome size

• Number of contigs

• N50 / L50

• GC content

• Presence of Ns

If you have many genomes and plan to submit them to NCBI or a Microbial Resource Announcement, you can extract QUAST results as a summary table using Python 🐍 Example Python script,quast_to_table.py, can be find at day 2 of https://github.com/jojyjohn28/whole-genome-sequencing-analysis

📈 GC Content Visualization (R)

GC content is a quick diagnostic for contamination or misassemblies. Eaxmple R script GC_vis.R also can be find at day 2 of https://github.com/jojyjohn28/whole-genome-sequencing-analysis

GC content

🔁 Genome Topology: Linear vs Circular

Determining topology is essential for:

• NCBI submission

• Genome announcements

• Biological interpretation

General rules:

• Short-read assemblies → treat as linear/draft

• Long-read Flye assemblies → may be circular

To validate circularity:

•Inspect Flye logs

• Check for overlap at contig ends

• Confirm no duplicated terminal regions More details and step by step instruction are published at : https://jojyjohn28.github.io/blog/genome-topology-and-genome-report/

🧠 Assembly Graph Inspection with Bandage

Bandage allows visual inspection of assembly graphs. Assembly graph visulization using bandage can be find at : https://jojyjohn28.github.io/blog/genome-visualization/

🧠 Key Takeaways from Day 2

● No single assembler fits all data ● Always evaluate assemblies with QUAST ● GC content is a powerful sanity check ● Long reads enable topology resolution ● Bandage helps see what statistics cannot

All batch scripts are available in the project repository. https://github.com/jojyjohn28/whole-genome-sequencing-analysis/day2

In Day 3, we will move from assemblies to taxonomy, phylogeny, and species delineation using GTDB-Tk, ANI, AAI, and dDDH.