De novo Genome Assembly Using Long Reads and Chromosome Conformation Capture.

The field of genome assembly merely exists as long as sequencers are not able to yield chromosome-level error-less sequencing reads for all species. It consists in reconstituting the original genome sequence from sequencing reads, with a final number of fragments matching the expected number of chromosomes. This process has been facilitated by the availability of longer and more accurate reads. At the incipit of genome assembly, Sanger sequencing reads Sanger et al (Proc Natl Acad Sci 74(12)spiepr A3B2 twbch ":":spiepr A3B2 twbch5463-5467, 1977) were already used to yield initial assemblies of different species, including the first human genome assembly International Human Genome Sequencing Consortium (Nature 409(6822):860-921, 2001). The higher throughput of second-generation sequencing, often called next-generation sequencing, democratized assemblies for a wider variety of species but brought assembly difficulties as the large datasets of short reads required long computational time, large memory resources, and yielded highly fragmented assemblies, with fragment numbers far over the expected number of chromosomes Metzker (Nat Rev Genet 11(1):31-46, 2010). Third-generation sequencing introduced long reads through the technologies of Oxford Nanopore Deamer et al(Nat Biotechnol 34(5):518-524, 2016) and Pacific BioSciences (PacBio) Eid et al (Science 323(5910):133-138, 2009). Long reads can reach several tens of kilobase, and up to hundreds of thousands of base pairs; although these reads initially had a low accuracy, recent developments to decrease the error rate below 1% Sereika et al (Oxford Nanopore R10.4 long-read sequencing enables near-perfect bacterial genomes from pure cultures and metagenomes without short-read or reference polishing. bioRxiv, 2021); Wenger et al (Nat Biotechnol 37(October):1155-1162, 2019) have additionally reduced the complexity of genome assembly.Chromosome-level assemblies have become a standard in genome assembly publications: they can be used for synteny analysis, finding chromosomal rearrangements, they have more complete gene sets, a better resolution of repetitive content, and fewer contaminants. The availability of long reads and Hi-C and the decreasing cost of sequencing have brought about many high-quality insect genome assemblies, with currently 6,194 assemblies published on GenBank (accessed on 20.11.2024). Some remarkable efforts have been put toward Lepidoptera genomes, for which 188 chromosome-level assemblies were generated by the Darwin Tree of Life Wright et al (Nat Ecol Evol 8(4):777-790, 2024). At a smaller scale, a highly contiguous assembly was obtained for the ant Camponotus pennsylvanicus using Nanopore reads with a budget of only 1000$.Traditionally, genome assemblies are collapsed, meaning that sets of homologous chromosomes are represented by a single consensus sequence Guiglielmoni et al (BMC Bioinf 22(1):303, 2021). This approach is most adequate for low-heterozygosity genomes, and variants are documented a posteriori. Along the advent of high-accuracy long reads, phased assemblies, in which all haplotypes are included, are becoming more common.The assembly process typically involves multiple steps to tackle the challenges posed by the characteristics of the genome: ploidy, heterozygosity, repetitiveness… Reads may need to be preprocessed to remove adapters, select the longest and/or most accurate reads, or correct them to further improve accuracy. The most essential part lies, of course, in the assembly step. Reads are overlapped to build an assembly graph, and ad hoc algorithms are applied to find a path giving the most faithful representation of the genome. Assemblers yield a set of contiguous sequences or contigs. The output should then be evaluated to decide whether the assembly has reached the highest contiguity, completeness, and correctness, and if not, which steps should be performed subsequently to increase quality.