Structural differences between the X and X' chromosomes from Bradysia coprophila, and the first assembled model sequence of the ~55 Mb X' long paracentric inversion region

The goal of the experiments and data shared through this BioProject and the associated research paper (Robert B. Baird, John M. Urban, Andrew J. Mongue, Kamil S. Jaron, Christina N. Hodson, Malte Grewoldt, Simon H. Martin, and Laura Ross, 2022; https://doi.org/10.1101/2022.11.24.517840) was to study the X' (X prime) chromosome in the dark-winged fungus gnat, Bradysia coprophila (a.k.a. Sciara coprophila). The male somatic Bradysia coprophila genome contains three autosomes (II, III, and IV) and the X chromosome, and was assembled previously with long reads and optical maps (Urban et al, 2021 ; https://doi.org/10.1186/s12864-021-07926-2 ; BioProject PRJNA291918 ; VSDI01000000 ), and subsequently scaffolded at chromosome-scale with Hi-C (Urban et al, 2022; https://doi.org/10.1101/2022.11.03.515061 ; BioProject PRJNA291918 ; VSDI02000000). However, there are two types of females: those that produce male offspring and are XX, and those that produce female offspring and are X'X. The previous male somatic genome assemblies did not include the X' chromosome and, until now, nothing was known about the X' at the sequence level. Previously, using cytogenetic methods, the X' chromosome was shown to differ from the X chromosome by a long paracentric inversion (Crouse, 1979). We set out to identify the breakpoints of this long paracentric inversion using DNA sequencing data from three groups of whole adult flies: XO males, XX male-producing females, and X'X female-producing females. The XO male and XX female data are used as negative controls (negative for the X' chromosome). The X'X female data is unique from the negative controls in containing data from the X' chromosome. We looked at patterns of discordantly-mapping paired-end reads (structural variation signals) from all three samples along the X chromosome in the somatic assembly (X, II, III, IV; VSDI02000000). The XO and XX samples have relatively few structural variation signals with respect to the X chromosome scaffold, and they are all also seen in the X'X sample. Conversely, the X'X sample contains a surprising amount of structural variation signal from paired reads discordantly-mapped to the X chromosome scaffold. This trend was also true when we compared structural variation signals from XO and X'X PacBio long read samples. The patterns of discordance and split read alignments suggest that the structural difference between the X and X' chromosomes is more complex than a single, long paracentric inversion, and is consistent with a more complicated set of nested rearrangements and/or transpositions. The three genotypes (XO, XX, X'X) were also used to look at single-nucleotide variation (SNV) along the X chromosome. After subtracting out variants shared by all samples, the density of X'X-specific variants along the X showed several blocks with different SNV densities, again suggesting multiple rearrangements that occurred sequentially over evolutionary time. However, Hi-C data generated from whole X'X female adults seemed to have strong SV signals only for the long paracentric inversion breakpoints. Finally, we identified kmers that were specific to the X'X sample (but more prevalent than error kmers), and used the X'-specific kmers to identify X'-specific reads to aid in assembling X' contigs. The X' contigs were scaffolded along the available X chromosome scaffold (Urban et al, 2022; https://doi.org/10.1101/2022.11.03.515061 ; BioProject PRJNA291918 ; VSDI02000000), demonstrating that they made up sequence strictly in between the long paracentric inversion breakpoints. Gaps in the X' inversion scaffold were filled with X'X PacBio long reads that preferentially mapped to the X' sequence in the presence of the X sequence. The X' contigs and scaffold were annotated and used to study the degradation of the X' chromosome/genes relative to the X chromosome. See the study for more details (Robert B. Baird, John M. Urban, et al, 2022; https://doi.org/10.1101/2022.11.24.517840).