Population genomic sequencing of Coccidioides fungi reveals recent hybridization and transposon control

Abstract

We have sequenced the genomes of 18 isolates of the closely related human pathogenic fungi Coccidioides immitis and Coccidioides posadasii to more clearly elucidate population genomic structure, bringing the total number of sequenced genomes for each species to 10. Our data confirm earlier microsatellite-based findings that these species are genetically differentiated, but our population genomics approach reveals that hybridization and genetic introgression have recently occurred between the two species. The directionality of introgression is primarily from C. posadasii to C. immitis, and we find more than 800 genes exhibiting strong evidence of introgression in one or more sequenced isolates. We performed PCR-based sequencing of one region exhibiting introgression in 40 C. immitis isolates to confirm and better define the extent of gene flow between the species. We find more coding sequence than expected by chance in the introgressed regions, suggesting that natural selection may play a role in the observed genetic exchange. We find notable heterogeneity in repetitive sequence composition among the sequenced genomes and present the first detailed genome-wide profile of a repeat-induced point mutation (RIP) process distinctly different from what has been observed in Neurospora. We identify promiscuous HLA-I and HLA-II epitopes in both proteomes and discuss the possible implications of introgression and population genomic data for public health and vaccine candidate prioritization. This study highlights the importance of population genomic data for detecting subtle but potentially important phenomena such as introgression.

Figure 1.

Sliding-window analysis of species divergence (F_st). Each circle represents F_st calculated for a nonoverlapping 5-kb window. High F_st values indicate strong local divergence between C. immitis and C. posadasii.

Figure 2.

Geographic profile of introgression from a deeper population sample. The Southern California population of C. immitis, which is more proximal to the range of C. posadasii, exhibits a greater incidence of introgression near the MEP4 locus than the Central California population (χ² test; P = 0.01). The approximated ranges of C. immitis and C. posadasii are indicated with gray shading and hashing, respectively.

Figure 3.

Dinucleotide mutational profile of repetitive and nonrepetitive sequence, normalized by dinucleotide abundance. The vertical axis represents the fraction of SNP-callable locations that are polymorphic (averaged across all isolates for a species), relative to the appropriate reference genome (RS for C. immitis and RMSCC_3488 for C. posadasii). SNP rates for all dinucleotides are higher in repetitive sequence and also in C. posadasii isolates. In addition, CpG and CpC show the highest mutational rates.

Figure 4.

Mutational profile across 23 gypsy element insertions. The C. immitis RS genome contains 23 full-length copies of a particular family of gypsy elements, 19 of which have substantial levels of modification at CpG residues. Analyzing only those locations that contain CpG in any of the 23 elements, this figure illustrates the two kinds of modifications seen. The green and red data points illustrate the almost complete modification of CpG to CpA (or TpG if on the other strand). There are no elements with intermediate levels of CpG content. The yellow data points illustrate the modification of CpA dinucleotides to a TpA occurring at a much more gradual pace, perhaps by a distinct process from that which makes the initial, high-efficiency CpG modifications.