A Kiss of Deep Homology: Partial Convergence in the Genomic Basis of Hypertrophied Lips in Cichlid Fish and Human Cleft Lip

Abstract

The genomic loci generating both adaptive and maladaptive variation could be surprisingly predictable in deeply homologous vertebrate structures like the lips. Variation in highly conserved vertebrate traits such as the jaws and teeth in organisms as evolutionarily disparate as teleost fishes and mammals is known to be structured by the same genes. Likewise, hypertrophied lips that have evolved repeatedly in Neotropical and African cichlid fish lineages could share unexpectedly similar genetic bases themselves and even provide surprising insight into the loci underlying human craniofacial anomalies. To isolate the genomic regions underlying adaptive divergence in hypertrophied lips, we first employed genome-wide associations (GWAs) in several species of African cichlids from Lake Malawi. Then, we tested if these GWA regions were shared through hybridization with another Lake Malawi cichlid lineage that has evolved hypertrophied lips seemingly in parallel. Overall, introgression among hypertrophied lip lineages appeared limited. Among our Malawi GWA regions, one contained the gene kcnj2 that has been implicated in the convergently evolved hypertrophied lips in Central American Midas cichlids that diverged from the Malawi radiation over 50 million years ago. The Malawi hypertrophied lip GWA regions also contained several additional genes that cause human lip-associated birth defects. Cichlid fishes are becoming prominent examples of replicated genomic architecture underlying trait convergence and are increasingly providing insight into human craniofacial anomalies such as a cleft lip.

Keywords: animal model; evolutionary medicine; genome resequencing; human birth defects.

Fig. 1.

Phylogenetic distribution of the hypertrophied lip phenotype in African Lake Malawi cichlids. Tree modified from Masonick et al. (2022) with hypertrophied lip lineages denoted by thickened branches; this phenotype likely evolved at least twice among the Lake Malawi cichlid fauna (stars). Hypertrophied lips appear to have evolved once in the primarily algae-scraping and rock-dwelling mbuna (A. labrosus) and another time among the largely sand-dwelling non-mbuna. Identifying the genomic basis of this phenotype will illuminate whether adaptive introgression has led to the peculiar phylogenetic distribution of this phenotype within Malawi and clarify whether aspects of the genomic basis of this trait are convergently shared with distantly related hypertrophied lip cichlids in the Neotropics as well as human craniofacial and lip anomalies. The phylogenetic position of M. mola, a thin-lipped species also included in our data set for GWA, is indicated.

Fig. 2.

GWA mapping of the hypertrophied lip phenotype in Lake Malawi cichlids. (a) Manhattan plot arranged by chromosome number (1–22). Numerous high signals of association (red line denotes 233 SNPs surpassing the threshold of −log₁₀(P) = 7.301) were detected across the genome. (b–d) Genomic windows with gene annotations of candidate lip loci on chromosomes 8, 12, and 13 are highlighted. Several genes occurring around identified peaks of high association have been found to be associated with the hypertrophied lip phenotype in other cichlid studies. (b) For instance, kcnj16 and kcnj2a are genes encoding for inward rectifier potassium channels and underlie the convergently evolved hypertrophied lip phenotype in Midas cichlids (Kautt et al. 2020). (c) The genes ggt5a and lrrc75ba and several other protein-coding genes are dispersed around a major peak of association located on chromosome 12. (d) In addition to these, col21a1, a gene implicated in the development of cleft lips in humans, overlaps with one of the strongest regions of genomic association identified. Supplementary table S5, Supplementary Material online, provides a list of annotated genes located within 50 kbp of these highest-associated SNPs.

Fig. 3.

Evaluation of introgressive hybridization of hypertrophied lip–associated SNPs in Lake Malawi cichlids along chromosome 13. The genomic position of the highly associated hypertrophied lip gene col21a1 is indicated by the green bar in graphs a–c. (a) We investigated whether GWA regions identified in the non-mbuna hypertrophied lip clade showed high associations within the lip size. Highly associated SNPs (P < 5e⁻⁸) are distinguished by the red line. (b) Genomic regions were examined using Dinvestigate to pinpoint regions showing higher than expected levels of introgression (high f_dM statistic calculated in sliding windows of 250 SNPs) in A. labrosus compared with other mbuna. The blue line denotes no introgression (f_dM = 0.0) whereas values strongly departing from 0.0 are more consistent with gene flow between the non-mbuna hypertrophied lip species and either A. labrosus (f_dM > 0.1) or the mbuna sans A. labrosus (f_dM < −0.1). (c) Additionally, we used a topology-weighting analysis (TWISST) to investigate whether the hypertrophied lip mbuna and non-mbuna were more likely to be monophyletic at putative GWA regions than expected by chance. Topology 1 (blue) reflects the accepted species tree (the mbuna (M) and A. labrosus (A) are monophyletic), Topology 2 (red) indicates shared phylogenetic signal between A. labrosus and non-mbuna haplochromines with hypertrophied lips (H), and Topology 3 (yellow) represents shared history between the mbuna (sans A. labrosus) and non-mbuna hypertrophied lip species. The non-Lake Malawi H. bloyeti was used as an outgroup (O). The raw topological frequency weighting was estimated across nonoverlapping windows of 250 SNPs spanning chromosome 13 and is plotted here. In general, high GWA regions in the hypertrophied lip non-mbuna (fig. 2a) do not show clear patterns of introgression into the Malawi mbuna hypertrophied lip species A. labrosus based on the Dinvestigate results; however, col21a1 is nested within a region that is enriched for Topology 2 according on the TWISST analysis.

Fig. 4.

Various hypertrophied lip–associated loci aligned to our chromosome-level GWA results. Previous studies that employed GWA mapping (red) in Midas cichlids (Kautt et al. 2020), QTL analysis (blue) of a hypertrophied lip species from Lake Victoria (Henning et al. 2017), and comparative transcriptomic gene expression (green) of lips in various lineages of cichlids (Colombo et al. 2013; Manousaki et al. 2013; Lecaudey et al. 2021), along with the GWA mapping conducted herein using hypertrophied lip cichlid species from Lake Malawi (Manhattan plot), have identified a number of loci that overlap with genes linked with craniofacial and lip anomalies in humans (black). The Lake Victoria QTL regions (supplementary table S6, Supplementary Material online) are shown primarily to highlight how they demarcate wide intervals on the chromosomes that in some cases could overlap peaks of association reported here for Malawi cichlids. Also, these QTLs, in conjunction with our GWA mapping, are consistent with a polygenic basis to hypertrophied lips in East African cichlids.