Karyotype depends on sperm head morphology in some amniote groups

The karyotype of an organism is the set of gross features that characterize the way the genome is packaged into separate chromosomes. It has been known for decades that different taxonomic groups often have distinct karyotypic features, but whether selective forces act to maintain these differences over evolutionary timescales is an open question. In this paper we analyze a database of karyotype features and sperm head morphology in 103 mammal species with spatulate sperm heads and 90 sauropsid species (birds and non-avian reptiles) with vermiform heads. We find that mammal species with a larger head area have more chromosomes, while sauropsid species with longer heads have a wider range of chromosome lengths. These results remain significant after controlling for genome size, so sperm head morphology is the relevant variable. This suggest that post-copulatory sexual selection, by acting on sperm head shape, can influence genome architecture.

One mammal species in the database that might qualify as an ambiguous "edge" case is Manis pentadactyla (pangolin), as its sperm head is longer and more cylindrical than typical (Chang et al., 2020).However, the apex of the head is still flattened, and the ultrastructure of the nucleus and acrosome is otherwise typical of spatulate mammals, so it was included in our analysis.
Sauropsids.Most bird and non-avian reptile species have an approximately cylindrical, conical, or vermiform head shape, with a relatively small acrosomal complex at the distal end.
From the sauropsids, we only excluded order Passeriformes (passerine birds), as these have a more complex helical or spindle-shaped head with an acrosome that can be longer than the nucleus, and a membranous structure that winds in a helix around the head (Støstad et al., 2018).

II. Use of TimeTree
Phylogenetic trees for Mammalia and Sauropsida were generated using TimeTree version 5 (Kumar et al., 2022).To generate a complete phylogeny, TimeTree automatically made congeneric substitutions for Giraffa camelopardalis, Miniopterus schreibersii, Eulemur fulvus, Jacana jacana, Phoenicopterus ruber, and Gyps fulvus.Similarly, we substituted Teius teyou for Teius oculatus, Eptesicus fuscus miradorensis for Eptesicus fuscus, and Crotallus molossus molossus for Crotallus molossus in the input list for TimeTree.In no case did our database include a second species in the same genus that might confound the use of a substitution.TimeTree produces a phylogenic tree in Newick format, which was subsequently imported into R. Plots of the phylogenetic trees are appended to this Supplement.

III. OLS regression details.
Lmin.We conducted a simultaneous ordinary least squares (OLS) regression of Lmin on genome size C, chromosome dispersity K, and chromosome number n:  Generated using TimeTree v 5.0 (Kumar et al., 2022).
, panels A and D, N = 103 (all mammal species in database).The two rows with head length L do not correspond to figures in the main text.They are provided to demonstrate the qualitative similarity of area and length as independent variables.In the main text we prefer to use area for spatulate heads since it is more directly related to nuclear shape changes during spermiogenesis, as described in the Discussion section.The two rows showing n vs. L and K vs. L with N = 55 do not correspond to figures in the main text.They are provided for comparison with the N = 90 data set above.The exponent β of the K vs. L regression (highlighted with an asterisk below) is not significant in the IV. PGLS regression details.Gross morphology.We tested the importance of sperm head morphology by introducing the parameter s (s = 0 for mammals, s = 1 for sauropsids).The PGLS regression is for log10(n) = α + β s and log10(K) = α + β s.Reported p-values are from 2-sided t-tests with the null hypothesis β = mammal entries).The two rows showing n vs.A and K vs.A with N = 87 do not correspond to figures in the main text.They are provided for comparison with the N = 103 data set above.