Goal: Reconstruct the sequences of a complete genome, or as close as possible to the complete genome, from sequences of DNA fragments (the "reads").

Genome assembly consists of aligning and reconstructing these fragments to form a continuous sequence (that of the chromosomes) or a set of contiguous sequences (called contigs or scaffolds).

• Step 1: Gather information about the target species : location, overlap with Indigenous People or local communities territories, expected genome size, ploidy, micro-chromosomes, organelles, etc. What type of data is already available (reference gneomes, raw data, etc) : INSDC (ENA, NCBI, DBNJ)
  
  
• Step 2: If appropriate (depending on species and sampling location), engage with local communities, Indigenous people and/or local scientists who may have Traditional Ecological Knowledge (TEK) associated to the species. Obtain appropriate consent and permits (Nagoya protocol) before sampling.
  
  
• Step 3: Build a broad community of collaborators for the project, if possible.
  
  
• Step 4: Select the best possible DNA source and an optimal extraction procedure - Sampling is THE key step
  
  
• Step 5: Choose an appropriate sequencing technology
  
  
• Step 6: Sequence and assemble!

• Many DNA extraction protocol are available for a wide range of species/taxa (VGP, Darwin Tree of Life, Nanopore, PacBio, etc)
  
  
• Keep DNA samples from the same individual in case of library preparation or sequencing failure, need more coverage, new sequencing technology, etc
  
  
• Use a single individual and sequence a haploid, a highly inbred diploid organism, or an isogenic individual

Contig: a contiguous sequence in an assembly. A contig does not contain long stretches of unknown sequences (aka assembly gaps). The contig is usually generated using the long-reads data.
Scaffold: a sequence consists of one or multiple contigs connected by assembly gaps of typically inexact sizes. A scaffold is also called a supercontig, though this terminology is rarely used nowadays. Usually, scaffolds are generated using the Hi-C data
Assembly: a set of contigs or scaffolds.

Hi-C: Capturing interactions between different parts of a genome by measuring the physical proximity of DNA segments in the nucleus:
Binding of closely interacting DNA regions.
DNA is digested, labeled, and joined using ligations to create hybrid fragments.
These fragments are sequenced to reveal which parts of the genome were spatially close, even if they are distant in terms of linear sequence.

Hi-C allows for the transition from the assembly of fragmented contigs to:

• a high-quality assembly, with scaffolds
• a whole chromosome assembly

This mainly depends on the quantity and quality of DNA as well as the cost of the experiment but many parameters need to be considered before performing an NGS experiment:

• Short versus long reads or both
• Read length
• Read quality/error rate
• Genome read coverage/depth : Number of unique reads that include a given nucleotide in the reconstructed sequence. -30X coverage mean that, on average, each nucleotide in the genome is covered bu 30 reads.
  • Coverage = (read count * read length ) / total genome size
• Library preparation
• Available technology
• Downstream applications

Sequencing technology for assembly:

• PacBio Hifi: long reads (up to 20kb)
• Nanopore: long reads and ultra-long read (up to 100kb)
• Illumina or MGI: short reads (up to 2x250bp) with high quality reads. Sequencing bias with AT/GC rich regions

Sequencing technology for scaffolding:

• Hi-C: restriction enzyme fragmentation (single, multiples sites or DNAse). Need huge amount of coverage. Providers : Arima Genomics, Phase Genomics, Dovetail Genomics

• Optical mapping: technique to physically locate specific enzymes restriction sites or sequence motifs to produce DNA sequence fingerprints. Providers : BioNano, BGI

• Mate pair (deprecated)

• BAC/YAC/Fosmids (deprecated)

Typical sequencing strategies - EBP (Earth Biogenome Project) recommendation -

• Long-reads (PacBio HiFi or ONT) : 15x per haplotype
• Hi-C data (Arima / Illumina) Polishing is no longer necessary or recommended

Different level of assembly exist today :

• Contig Assembly
  
• Scaffold assembly
  
• Chromosome-level assembly : When the number of scaffolds is the number of expected chromosomes, it means that 1 chromosome = 1 scaffold, and no large string of sequence is unlocated.

It can still contain gaps in between the scaffolds (shown as "NNNNNNNN" in the assembly)

• T2T (Telomere-to-Telomere) : Assembly without any gap (a chromosome level assembly without "NNNNN" sequences)

Haplotig: a contig that comes from the same haplotype. In an unphased assembly, a contig may join alleles from different parental haplotypes in a diploid or polyploid genome.
Primary assembly: a complete assembly with long stretches of phased blocks.
Alternate assembly: an incomplete assembly consisting of haplotigs in heterozygous regions. An alternate assembly always accompanies a primary assembly. It is not useful by itself as it is fragmented and incomplete.
Haplotype-resolved assembly: sets of complete assemblies consisting of haplotigs, representing an entire diploid/polyploid genome.

To be successful, you must have sufficient computing resources (CPUS, RAM, walltime and storage).

• The resources needed are different for each step:
  • Assembly
  • Annotation
  • Other analysis tools
    
• For genome assembly:
  • Running times and RAM increase with data type and amount
  • More data for large genomes, increase runtime/RAM/Storage
  • Most of tools run on a single node: they are parallelized but not distributed
    
• For genome annotation:
  • Mapping/alignment of external data (RNA-seq, proteins) can be parallelized and distributed
  • Annotation process can be parallelized and distributed

FASTA: a text-based format for representing either nucleotide sequences or amino acid (protein) sequences, in which nucleotides or amino acids are represented using single-letter codes.

FASTQ: a text-based format for storing both a biological sequence (usually nucleotide sequence) and its corresponding quality scores (Phred). Both the sequence letter and quality score are each encoded with a single ASCII character for brevity. It's the standard sequencing output for Illumina and MGI sequencers.

N50: given a set of sequences of varying lengths, the N50 is defined as the length L of the shortest contig for which longer and equal length contigs cover at least 50% of the assembly.
L50: given a set of sequences of varying lengths, the L50 is defined as count of smallest number of sequences whose length sum makes up 50% of the assembly.
N50 describes a sequence length whereas L50 describes a number of sequences.

Example:

• Genome size = 100
• Sequence sorted by size list L = (25, 10, 10, 8 , 7, 7 , 6 , 5, 5, 5, 5, 3, 2, 2 ) = 100
• 50% of the total length is contained within sequences of at least 8bp: 25 + 10 + 10 + 8 ≥ 50