Extended Data Fig. 2: Chromosome lengths and comparison of assembly completeness in nonexonic genomic regions.

a, Comparison of scaffold lengths and chromosome lengths that were estimated from published karyotype images of M. molossus, M. myotis and R. aegyptiacus. b, To assess completeness in nonexonic genomic regions, we determined how many of 197 nonexonic ultraconserved elements (UCEs)²⁰ align at ≥ 85% identity to the human sequence. UCEs are highly conserved among mammals⁸⁸ and are expected to be present in complete assemblies. Bar charts show the number of detected UCEs that align at these stringent parameters. As expected, the vast majority of UCEs were detected in all assemblies. UCEs not detected are separated into those that are missing owing to assembly incompleteness and those that exhibit real sequence divergence. In the bat genomes we report, no UCEs were missing owing to assembly incompleteness. Instead, one to three UCEs were not detected in our Myotis and Pipistrellus assemblies because the UCE sequences are more than 85% diverged (Supplementary Fig. 1). Human and mouse are not shown here because both genomes were used to define ultraconserved elements²⁰. For cow and cat, we also compared new assemblies (bosTau9 and felCat9) that recently became available. c, Example of a UCE that is not fully present in the assemblies of cow (bosTau8), cat (felCat8) and dog (canFam3) because of assembly gaps. UCSC genome browser screenshot shows a multiple genome alignment of mammals of the locus around UCE.157 (highlighted) and pairwise chains of co-linear alignments (blocks represent local alignments, double lines represent unaligning sequence and single lines represent deletions). The top-level pairwise alignment chains between human (reference species) and cow, cat and dog show that UCE.157 only partially aligns (cow bosTau8 and dog canFam3) or does not align at all (cat felCat8). The unaligning region overlaps an assembly gap in all three cases, indicating that the UCE sequence is not present because of assembly incompleteness. Indeed, the UCE is entirely present in more-recent assemblies of cow (bosTau9) and cat (felCat9). Furthermore, the alignment chains of the dingo—a close relative of the dog—show that the dingo assembly also contains the entire UCE.157. d, Example of a UCE that shows real sequence divergence in Pipistrellus bats. Dots in the alignment represent nucleotides that are identical to the human sequence shown at the top. Compared to other bats, Pipistrellus kuhlii shows an increased number of mutations in this UCE sequence; however, M. myotis also shows an increased number of mutations. Because most mutations are shared between P. kuhlii and Pipistrellus pipistrellus, base errors in the assembly are highly unlikely to account for the increased sequence divergence.