What is Genome Assembly?

Genome assembly involves piecing together short DNA sequence reads generated by sequencing platforms (e.g., Illumina, PacBio, Oxford Nanopore) into longer, contiguous sequences called contigs. This can be performed as:

• De Novo Assembly: Without a reference genome.
• Reference-Guided Assembly: Using a reference genome to guide the assembly process.

Step 1: Preparing Your Data

Before starting the assembly, ensure that your raw sequencing data is high quality.

1. Input Data

   • Short Reads: Illumina sequencing generates short, accurate reads ideal for scaffolding.
   • Long Reads: PacBio and Nanopore sequencing provide long reads for resolving repetitive regions.

2. Quality Control (QC)
   Use tools like FastQC or MultiQC to assess the quality of your reads:

   fastqc reads.fastq multiqc .

   Look for issues like low-quality bases, adapter contamination, or overrepresented sequences.

3. Read Trimming and Filtering
   Trim low-quality bases and adapters using Trimmomatic or Cutadapt:

   trimmomatic PE reads_R1.fastq reads_R2.fastq trimmed_R1.fastq trimmed_R2.fastq \ ILLUMINACLIP:adapters.fa:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:20 MINLEN:36

Step 2: Choosing an Assembly Strategy

Select an assembly strategy based on your data type:

• Short-Read Assemblers:

  • SPAdes: Popular for microbial genomes.
  • Velvet: Fast for smaller genomes.

• Long-Read Assemblers:

  • Canu: Ideal for long-read datasets.
  • Flye: Versatile for small and large genomes.

• Hybrid Assemblers:

  • MaSuRCA: Combines short and long reads.
  • Unicycler: Optimized for bacterial genomes.

Step 3: Running the Assembly

3.1. SPAdes (Short-Read Assembly)

SPAdes is an excellent choice for small genomes, such as bacteria.

The output includes assembled contigs (contigs.fasta) and scaffolds (scaffolds.fasta).

3.2. Canu (Long-Read Assembly)

Canu is designed for high-error long reads from PacBio or Nanopore.

The output will be in canu_output/genome.contigs.fasta.

3.3. Hybrid Assembly with Unicycler

Unicycler combines short and long reads for improved assemblies.

Step 4: Assessing Assembly Quality

After assembly, evaluate its quality using the following tools:

1. QUAST
   QUAST generates assembly statistics, such as N50, genome size, and GC content:

   quast contigs.fasta -o quast_output

2. BUSCO
   BUSCO checks genome completeness by identifying conserved genes:

   busco -i contigs.fasta -o busco_output -l fungi_odb10 -m genome

3. Assembly Graph Visualization
   Visualize assembly graphs with Bandage:

   Bandage load assembly_graph.gfa

Step 5: Post-Assembly Steps

1. Polishing
   Improve assembly accuracy using tools like Pilon (for short reads) or Racon (for long reads).

   racon long_reads.fasta mapped_reads.sam contigs.fasta > polished_contigs.fasta

2. Scaffolding
   Link contigs into scaffolds using tools like SSPACE or Opera-LG if required.

3. Annotation
   Annotate the assembled genome using Prokka for prokaryotes or Maker for eukaryotes.

   prokka --outdir annotation_output --prefix genome contigs.fasta

Step 6: Sharing and Archiving

1. Submit to Public Repositories
   Share your assembly in databases like NCBI GenBank, ENA, or DDBJ.

2. Metadata Preparation
   Include detailed metadata for your submission, such as organism name, sequencing platform, and coverage.

Best Practices

• Always perform quality checks at each stage to ensure data integrity.
• Use multiple tools to cross-validate results when working with complex genomes.
• Document parameters and software versions for reproducibility.

Conclusion

Genome assembly is a powerful process that transforms raw sequencing data into a coherent representation of an organism’s genome. By following this step-by-step guide, you can successfully assemble genomes and uncover valuable biological insights. Whether you’re assembling a microbial genome or tackling the complexities of a eukaryotic genome, these tools and strategies will set you on the path to success.

We have developed panHiTE, a comprehensive and accurate pipeline for TE detection in large-scale population genomes. It has been successfully applied to hundreds of plant population genomes, demonstrating its effectiveness and scalability.

For detailed instructions, please refer to the panHiTE tutorial.

Address of the bookmark: https://github.com/CSU-KangHu/HiTE

“It’s like a subway map,” she said, “but for DNA.”

That idea is exactly what Sequence Tube Map brings to life.

Instead of showing a genome as one long, confusing sequence, it displays different genomic paths like routes on a metro map. Where sequences are shared, the paths travel together. Where genetic variation occurs, they branch into different routes and may join again later.

This makes complex genome graphs much easier to explore.

Maya could now see how different haplotypes moved through the same genomic region and where they took different paths. She could even view sequencing reads mapped onto the graph, helping her understand what was happening in regions with genetic variation.

What once looked like a tangled network now looked like a journey.

Sequence Tube Map turns the complexity of graph genomes into a visual story—making it easier for scientists to see how DNA can take different routes.

Read more at  https://github.com/vgteam/sequencetubemap 

and

https://academic.oup.com/bioinformatics/article/35/24/5318/5542397

Demo at https://vgteam.github.io/sequenceTubeMap/

A recent study published in Genome Biology (https://link.springer.com/article/10.1186/s13059-026-04280-y) addressed this challenge by introducing dicast (https://github.com/burgshrimps/dicast), a machine-learning approach for detecting structural variants in short-read sequencing data. Structural variants, such as large deletions, insertions, duplications, and inversions, can have significant biological and clinical implications, yet they are challenging to identify with short-read technologies. Because different detection methods frequently yield divergent results, researchers commonly employ consensus calling, considering a variant valid only if multiple tools detect it. While this approach reduces false positives, it relies on the potentially flawed assumption that the majority is always correct.

The researchers explored the impact of evaluating the supporting evidence for each variant, rather than simply tallying the number of algorithms that identified it. To establish a ground truth, they analyzed nine genomes using multiple sequencing technologies and 15 detection methods, initially identifying approximately 35 million potential variants. Through extensive filtering, evidence integration, and manual review of over 11,500 variants, they developed a robust benchmark comprising more than 236,000 structural variants. The findings underscored the complexity of the problem: short-read methods detected fewer than half of deletions and less than 10 percent of insertions, with performance declining markedly in repetitive genomic regions. In contrast, long-read technologies demonstrated superior detection capabilities. However, replacing the substantial volume of existing short-read data in clinical and research settings is not immediately feasible. Consequently, the researchers questioned whether short-read data might harbor more information than conventional analytical pipelines currently extract.

This line of inquiry led to the development of dicast. Rather than merely confirming agreement among multiple tools, dicast identifies patterns in sequencing data, including split and clipped reads, discordant read pairs, alignment characteristics, and the surrounding genomic context. An XGBoost machine-learning model evaluates which combinations of these signals are indicative of genuine structural variants. Thus, the approach shifts from tallying algorithmic consensus to interpreting the underlying evidence.

The researchers subsequently conducted a targeted evaluation by examining structural variants detected by only a single short-read tool, which are typically missed by consensus-based approaches. dicast successfully recovered approximately 81% of these single-caller deletions, insertions, and duplications. The signals for these variants were present in the data, but conventional filtering methods failed to integrate them effectively.

The utility of dicast was further demonstrated in cohorts with rare diseases, including congenital limb malformations, atrial fibrillation, and neuromuscular disorders. In one instance, dicast achieved a deletion recall rate of approximately 0.96, compared to 0.74 using consensus calling. The median number of variants requiring manual review was 29 per sample. Among 31 experimentally validated variants that standard filters would have missed, dicast identified 12, whereas consensus calling detected only one.

Overall, dicast identified approximately 20 percent more potential disease-causing deletions than consensus-based methods. While a 20 percent increase may appear modest, in clinical genomics such improvements can have significant practical implications. Missing a deletion may leave a case unresolved, whereas detecting a structural variant can provide critical diagnostic insights.

The study does not claim that machine learning has rendered short-read sequencing superior to long-read approaches. Instead, the results underscore the effectiveness of long-read sequencing for structural variant detection. However, dicast highlights a more nuanced perspective: substantial biological information may still be recoverable from the extensive short-read datasets already available.

The principal lesson extends beyond the detection of structural variants. For many years, bioinformatics pipelines have relied on threshold-based criteria, such as minimum coverage, quality scores, or support from multiple tools. While these rules are useful, biological phenomena do not always conform to rigid checklists; multiple weak signals, when considered collectively, can provide compelling evidence.

This perspective prompts consideration of the solitary variant: one algorithm identifies it, while several others do not. Traditional consensus techniques might have dismissed it, yet machine learning approaches evaluate the available evidence to determine whether the variant is plausible.

Occasionally, the most significant variant within a genome is the one that is almost universally overlooked.

Read more at https://link.springer.com/article/10.1186/s13059-026-04280-y

Link:

http://www.nytimes.com/2013/11/21/science/two-surprises-in-dna-of-boy-found-buried-in-siberia.html?_r=0

http://www.nature.com/nature/journal/vaop/ncurrent/full/nature12736.html

http://www.eva.mpg.de/genetics/bioinformatics/overview.html?Fsize=0%2C%20%40%2F%27

Kelso Group
Department of Evolutionary Genetics
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6
04103 Leipzig
Germany
Phone: +49 341 3550 500

Job:
http://www.eva.mpg.de/genetics/bioinformatics/jobs.html?Fsize=0%2C%2B%40

Job Description: We are interested in finding an excellent postdoc with interests in protein functional annotation, machine learning and computer grids. The position is open for 3.5 years at the Université Pierre et Marie Curie, in the heart of paris.

Research topic: Protein function annotation, multiple probabilistic models, domain architecture, machine learning, combinatorial optimization, computer grid.

Title: A novel integrative platform for large scale protein annotation that exploits a multitude of diversified probabilistic models in several protein signature databases.

We propose a novel integrated approach for large scale protein annotation that will exploit an unprecedented amount of genomic data as well as sophisticated machine learning techniques and combinatorial optimization approaches taking advantages of High Performance Computing (HPC) environments. The idea is to uncover as much as possible the evolutionary processes of protein sequences that took place throughout the whole tree of life and that affected the evolution of a protein family. We have already demonstrated in a previous work that the problem of functional annotation is inherent to the ability of uncovering such paths. Now, we shall extend this approach to large scale genome annotation by considering 11 different protein databases, constituted by about 10^9 protein sequences, and by producing a large pool of diversified probabilistic models coding for about 10^7 evolutionary protein pathways. Such models will be used to search for specific domains in genomes to be annotated. Our previous methodology needs to be fundamentally improved to deal with this large amount of biological data. In this project, we shall work on the algorithms to reduce the space of models and the search complexity, and we shall implement some important algorithmic changes towards the realization of a powerful integrated annotation tool.

Where: This project is run on the Laboratoire de Biologie Computationnelle et Quantitative UMR7238 CNRS-UPMC – Analytical Genomics team, headed by A.Carbone. It is co-advised with Pierre-Henri Wuillemin, Laboratoire d’Informatique de Paris 6 – Equipe DECISION.

Start date: September 1st, 2014
Contact Person: Alessandra Carbone
Contact: alessandra.carbone@lip6.fr

Main research topics:
- Comparative and Population Genomics of Plant Symbionts
- Parasite Genome Evolution
- Experimental Evolution of Microbial Symbionts and Parasites
- Phylogenomics of Early Branching Fungi

More at http://corradilab.weebly.com/