- Contigs, scaffolds, and chromosome-level assemblies are fundamental concepts in genome assembly and genomic analysis. They describe different levels of continuity and organization in reconstructed DNA sequences. Genome sequencing produces millions or billions of individual DNA reads, which must be computationally assembled to reconstruct the original genome. Because sequencing reads cover only portions of the genome, bioinformatics algorithms identify overlaps, shared sequence patterns, and other relationships to combine them into longer sequences. The resulting assembly may consist of contigs, scaffolds, or chromosome-scale sequences, depending on the sequencing technologies, assembly methods, and supporting evidence available.
- A contig, short for contiguous sequence, is a continuous DNA sequence assembled from overlapping sequencing reads without internal gaps. It represents a region of the genome for which the nucleotide sequence has been reconstructed continuously. Contigs are generated when sequencing reads overlap sufficiently to establish their relative positions and orientations. Assembly algorithms combine these reads into consensus sequences, resolving local overlaps and correcting sequencing errors where possible. A contig may represent a small genomic region, a complete microbial chromosome, or part of a much larger chromosome, depending on genome complexity and sequencing coverage.
- Contig formation is influenced by several factors, including read length, sequencing accuracy, sequencing depth, genome size, repetitive DNA, heterozygosity, and the performance of assembly algorithms. Short reads may provide high nucleotide accuracy but often cannot span long repetitive regions. Long reads can extend across larger genomic regions and help connect sequences that would otherwise remain fragmented. Insufficient sequencing coverage may prevent the assembler from identifying reliable overlaps, while sequencing errors can create false overlaps or break otherwise continuous sequences. Repetitive regions are particularly challenging because similar sequences may occur at multiple genomic locations, making it difficult to determine the correct arrangement of reads.
- A genome assembly containing many short contigs is generally more fragmented than one containing fewer, longer contigs. However, contig length alone does not determine assembly accuracy. A long contig may contain incorrect joins or collapsed repeat regions, while shorter contigs may represent accurately reconstructed genomic sequences. Assembly evaluation therefore considers both continuity and correctness. Metrics such as contig N50, total assembly length, number of contigs, read-mapping consistency, and gene completeness provide complementary information about assembly quality.
- A scaffold is a larger assembly structure created by connecting two or more contigs in an estimated order and orientation. Unlike a contig, a scaffold may contain gaps between its component contigs. These gaps are commonly represented by consecutive unknown nucleotide symbols, usually N, because the exact intervening sequence has not been reconstructed. Scaffolding uses additional information from sequencing libraries and other genomic technologies to determine which contigs are likely to belong together, how they should be oriented, and approximately how far apart they are.
- For example, a scaffold may contain three contigs arranged in a specific order, with unknown nucleotide regions between them. The contigs provide the known DNA sequence, while the gaps represent estimated intervals whose exact sequences remain unresolved. Scaffolding improves the structural continuity of an assembly, but it does not necessarily recover the missing DNA. A scaffold can therefore be much longer than any individual contig without representing one completely reconstructed continuous nucleotide sequence.
- Paired-end sequencing is a common source of information for scaffolding. In paired-end sequencing, both ends of a DNA fragment are sequenced, producing two reads with an expected orientation and approximate separation. If the two reads map to different contigs, their positions and orientations can provide evidence that the contigs are adjacent or belong to the same genomic region. The expected insert-size distribution helps estimate the distance between them. Mate-pair sequencing provides longer-range linking information and can help connect contigs separated by larger repetitive regions. However, both methods can produce ambiguous links when reads map to repeated sequences or when library preparation introduces chimeric fragments.
- Long-read sequencing technologies provide additional opportunities for scaffolding. PacBio HiFi and Oxford Nanopore reads can span long genomic regions and connect contigs across repetitive sequences, structural variations, and other complex regions. A long read that aligns consistently to two contigs can provide evidence for their relative order and orientation. In some cases, long reads can also span the entire interval between contigs, allowing the missing sequence to be reconstructed rather than represented as an unknown gap. The effectiveness of long-read scaffolding depends on read length, read accuracy, sequencing coverage, and the uniqueness of the flanking regions.
- Hi-C sequencing is particularly useful for organizing scaffolds into chromosome-scale structures. Hi-C measures physical interactions between DNA regions that are spatially close within the nucleus. Regions located on the same chromosome often show characteristic interaction patterns, which can be used to group and order contigs or scaffolds. Hi-C data are widely used in plant and animal genome projects because they can support chromosome-level assemblies even when individual sequencing reads cannot span entire chromosomes. However, Hi-C provides information about spatial proximity rather than direct nucleotide continuity, so its results should be validated carefully.
- Optical mapping and genetic linkage maps can provide further evidence for scaffold organization. Optical mapping identifies patterns of labelled sites along long DNA molecules and compares them with patterns predicted from assembled sequences. This information can help identify large structural differences and support the ordering and orientation of scaffolds. Genetic linkage maps use recombination information from populations or experimental crosses to estimate the relative positions of genetic markers. They can help assign scaffolds to chromosomes and determine their order, although their resolution depends on marker density and the available genetic information.
- A chromosome-level assembly is a genome assembly in which sequences have been organized into chromosome-scale structures. Ideally, each chromosome is represented by a single continuous sequence extending from one end to the other. In practice, chromosome-level assemblies may still contain gaps, unresolved repetitive regions, or regions with uncertain sequence information. The term describes the level of large-scale genomic organization and should not automatically be interpreted as meaning that every nucleotide has been completely reconstructed.
- Chromosome-level assembly generally requires more than conventional short-read assembly. Short-read data can provide accurate sequence information, but their limited length often makes it difficult to resolve long repeats and large structural relationships. Long-read sequencing helps reconstruct extended genomic regions, while Hi-C, optical mapping, and genetic maps provide long-range information for chromosome organization. A combination of these technologies can produce more complete and reliable chromosome-scale assemblies.
- The distinction between contigs, scaffolds, and chromosome-level assemblies can be understood as a progression in genome organization. Contigs represent continuous reconstructed sequences. Scaffolds connect contigs using estimated relationships and may contain unresolved gaps. Chromosome-level assemblies organize long sequences into structures corresponding to individual chromosomes. This progression does not necessarily occur in a strictly linear manner because modern assemblers may generate chromosome-scale sequences directly from long reads and other supporting data. Nevertheless, the three terms describe important differences in sequence continuity, structural organization, and assembly completeness.
- The following simplified representation illustrates the relationship:
CONTIGS
Contig 1 Contig 2 Contig 3
[ATGCTA] [GGTACC] [TACGGA]
SCAFFOLD
[ATGCTA]—NNNNNN—[GGTACC]—NNNNNN—[TACGGA]
Contig 1 Contig 2 Contig 3
CHROMOSOME-LEVEL ASSEMBLY
[ATGCTA + reconstructed sequence + GGTACC
+ reconstructed sequence + TACGGA]
- In this illustration, the scaffold connects the contigs but retains unknown regions. The chromosome-level representation shows the intended organization after the missing sequences and larger genomic relationships have been resolved. Real chromosome-level assemblies can still contain unresolved gaps, and the example is a conceptual representation rather than a complete biological sequence.
- Assembly continuity is commonly assessed using N50 and L50 statistics. N50 is the sequence length at which 50% of the total assembly length is contained in sequences of that length or longer. L50 is the minimum number of the longest sequences required to account for at least 50% of the total assembly length. These statistics can be calculated separately for contigs and scaffolds. Scaffold N50 is often larger than contig N50 because scaffolding joins multiple contigs into longer structures, but this increase does not necessarily indicate improved nucleotide accuracy.
- For example, consider an assembly containing five contigs with lengths of 100 kb, 80 kb, 60 kb, 40 kb, and 20 kb. The total assembly length is 300 kb. The contigs are arranged from longest to shortest, and the first two contain 180 kb, which is more than half of the total assembly length. The N50 is therefore 80 kb, and the L50 is 2. This example illustrates how N50 and L50 describe assembly continuity rather than the correctness of individual sequences.
- Genome completeness is another important measure. BUSCO evaluates the presence of expected conserved single-copy orthologous genes from a selected evolutionary lineage. A high BUSCO completeness score indicates that many expected conserved genes are present in the assembly, but it does not prove that the genome is structurally correct or that repetitive regions have been resolved. K-mer-based tools such as Merqury can evaluate sequence completeness and consensus accuracy by comparing k-mers in sequencing reads with those in the assembly. QUAST provides assembly statistics and can support reference-based comparisons when a suitable reference is available.
- Several software tools are used to generate contigs, scaffolds, and chromosome-level assemblies. SPAdes, MEGAHIT, ABySS, and other de Bruijn graph assemblers are commonly used for short-read assembly. Flye, Canu, Hifiasm, Shasta, and Verkko support long-read assembly using different algorithms and data types. Unicycler is widely used for microbial genome assembly and can combine short- and long-read evidence. Scaffolding tools include SSPACE, RagTag, YaHS, SALSA2, and other methods designed for different types of linking information. QUAST, BUSCO, and Merqury support assembly evaluation, while Bandage can be used to inspect assembly graphs.
- A simple workflow for generating contigs from short reads may use SPAdes. An illustrative command is:
spades.py \
-1 reads_R1.fastq.gz \
-2 reads_R2.fastq.gz \
-o spades_output \
-t 16 \
-m 64
- The assembled contigs can be found in the SPAdes output directory, commonly in contigs.fasta. The scaffolds.fasta output contains scaffold sequences generated by the assembler, which may include gaps represented by unknown bases. The exact output and interpretation depend on the SPAdes version and assembly mode.
- For long-read assembly using Oxford Nanopore reads, Flye can be used to generate an initial assembly:
flye –nano-raw long_reads.fastq.gz \
–out-dir flye_output \
–threads 16 \
–genome-size 5m
- The estimated genome size should be adjusted for the organism being assembled. Other sequencing technologies, including PacBio HiFi, require the appropriate input mode. Flye produces contigs and assembly graph information that can be used for further inspection and evaluation.
- Hi-C data can be used to organize contigs into chromosome-scale scaffolds. For example, YaHS supports Hi-C-based scaffolding, although the required alignments and options depend on the data and software version. A simplified command may look like this:
- yahs contigs.fasta hic_alignments.bam
- The resulting scaffold sequences and associated assembly information should be evaluated using Hi-C contact maps and independent evidence where possible. Hi-C scaffolding can make substantial improvements to chromosome-scale organization, but incorrect joins and orientation errors may occur if the contact information is ambiguous.
- Reference-guided scaffolding is another approach. RagTag compares a query assembly with a reference genome and uses the relationship between them to propose scaffold order and orientation. An illustrative command is:
ragtag.py scaffold reference.fasta contigs.fasta \
-o ragtag_output \
-t 16
- Reference-guided scaffolding is particularly useful when a closely related, well-assembled reference genome is available. However, the resulting arrangement may incorrectly reflect the target genome if it contains structural differences, rearrangements, or substantial evolutionary divergence from the reference. Independent validation is important before treating the proposed arrangement as confirmed.
- The quality of contigs, scaffolds, and chromosome-level assemblies should be assessed using several complementary methods. QUAST can compare assembly statistics, including total length, contig and scaffold counts, N50, L50, and other indicators. BUSCO evaluates conserved gene completeness, while Merqury assesses k-mer-based completeness and consensus quality. Read mapping can help identify unsupported regions, unexpected coverage patterns, and possible misassemblies. Hi-C contact maps, optical maps, and genetic linkage information can provide additional evidence for chromosome-scale organization.
- A simplified QUAST evaluation command is:
quast.py assembly.fasta \
-o quast_results \
-t 16
- Researchers should evaluate both continuity and accuracy rather than relying on a single statistic. A large scaffold N50 does not guarantee that the scaffolds are correctly organized. Similarly, a chromosome-scale assembly may still contain unresolved repetitive regions or sequence errors. Assembly validation should therefore include nucleotide-level, gene-level, and structural assessments.
- Contigs, scaffolds, and chromosome-level assemblies are used in many areas of biological research. In microbial genomics, complete or near-complete bacterial chromosomes and plasmids support gene identification, comparative genomics, antimicrobial resistance analysis, and evolutionary studies. In human genomics, chromosome-scale assemblies improve the analysis of structural variation, duplicated genes, centromeric regions, and complex genomic loci. In animal and plant genomics, high-quality chromosome assemblies support genome annotation, comparative evolution, trait mapping, genetic improvement, and conservation research. In metagenomics, contigs and scaffolds are used to reconstruct microbial genomes from mixed communities, although organismal diversity and strain variation make chromosome-level reconstruction particularly challenging.
- The quality and interpretation of an assembly also depend on ploidy and biological variation. Diploid organisms contain homologous chromosome pairs, and their two haplotypes may differ at many positions. Some assemblers produce a primary assembly with alternative haplotypes represented separately, while haplotype-aware assemblers may attempt to reconstruct both chromosome copies. In polyploid organisms, multiple related chromosome sets create additional complexity. A chromosome-level assembly should therefore clearly document whether it represents one haplotype, a collapsed representation, or multiple separately reconstructed haplotypes.
- The terms contig, scaffold, and chromosome-level assembly should not be treated as direct indicators of biological completeness. A contig is a continuous sequence, but it may cover only a small genomic region. A scaffold provides larger-scale organization, but may contain unknown sequence gaps. A chromosome-level assembly provides chromosome-scale organization, but may still have unresolved sequence regions or structural uncertainties. Assembly reports should describe these characteristics clearly and distinguish sequence continuity from sequence accuracy and biological completeness.
- For reproducible genome research, it is important to preserve the original sequencing reads, assembled contigs, scaffolds, chromosome-scale sequences, assembly graphs, software versions, and command-line parameters. Researchers should document the sequencing technologies, scaffolding evidence, reference genomes, polishing procedures, and validation methods used. They should also report unresolved gaps, uncertain scaffold joins, and whether the assembly is haplotype-resolved. Such documentation allows other researchers to assess assembly quality and reproduce or extend the analysis.
- Contigs, scaffolds, and chromosome-level assemblies represent different stages and levels of genome reconstruction. Contigs provide continuous sequences, scaffolds connect contigs using long-range evidence, and chromosome-level assemblies organize genomic sequences into chromosome-scale structures. Each level provides different information and has different limitations. Understanding these distinctions is essential for interpreting genome assembly results, selecting suitable bioinformatics tools, evaluating assembly quality, and producing reliable genomic resources for research in genomics, medicine, agriculture, biotechnology, and evolutionary biology.