The Practical Haplotype Graph, a platform for storing and using pangenomes for imputation

MOTIVATION

Pangenomes provide novel insights for population and quantitative genetics, genomics, and breeding not available from studying a single reference genome. Instead, a species is better represented by a pangenome or collection of genomes. Unfortunately, managing and using pangenomes for genomically diverse species is computationally and practically challenging. We developed a trellis graph representation anchored to the reference genome that represents most pangenomes well and can be used to impute complete genomes from low density sequence or variant data.

RESULTS

The Practical Haplotype Graph (PHG) is a pangenome pipeline, database (PostGRES & SQLite), data model (Java, Kotlin, or R), and Breeding API (BrAPI) web service. The PHG has already been able to accurately represent diversity in four major crops including maize, one of the most genomically diverse species, with up to 1000-fold data compression. Using simulated data, we show that, at even 0.1X coverage, with appropriate reads and sequence alignment, imputation results in extremely accurate haplotype reconstruction. The PHG is a platform and environment for the understanding and application of genomic diversity.

AVAILABILITY

All resources listed here are freely available. The PHG Docker used to generate the simulation results is https://hub.docker.com/ as maizegenetics/phg:0.0.27. PHG source code is at https://bitbucket.org/bucklerlab/practicalhaplotypegraph/src/master/. The code used for the analysis of simulated data is at https://bitbucket.org/bucklerlab/phg-manuscript/src/master/. The PHG database of NAM parent haplotypes is in the CyVerse data store (https://de.cyverse.org/de/) and named /iplant/home/shared/panzea/panGenome/PHG_db_maize/phg_v5Assemblies_20200608.db.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Entering the second century of maize quantitative genetics

Maize is the most widely grown cereal in the world. In addition to its role in global agriculture, it has also long served as a model organism for genetic research. Maize stands at a genetic crossroads, as it has access to all the tools available for plant genetics but exhibits a genetic architecture more similar to other outcrossing organisms than to self-pollinating crops and model plants. In this review, we summarize recent advances in maize genetics, including the development of powerful populations for genetic mapping and genome-wide association studies (GWAS), and the insights these studies yield on the mechanisms underlying complex maize traits. Most maize traits are controlled by a large number of genes, and linkage analysis of several traits implicates a 'common gene, rare allele' model of genetic variation where some genes have many individually rare alleles contributing. Most natural alleles exhibit small effect sizes with little-to-no detectable pleiotropy or epistasis. Additionally, many of these genes are locked away in low-recombination regions that encourage the formation of multi-gene blocks that may underlie maize's strong heterotic effect. Domestication left strong marks on the maize genome, and some of the differences in trait architectures may be due to different selective pressures over time. Overall, maize's advantages as a model system make it highly desirable for studying the genetics of outcrossing species, and results from it can provide insight into other such species, including humans.

Technical note: Use of marker-based relationships with multiple-trait derivative-free restricted maximal likelihood

The widespread use of the set of multiple-trait derivative-free REML programs for prediction of breeding values and estimation of variance components has led to significant improvement in traits of economic importance. The initial version of this software package, however, was generally limited to pedigree-based relationships. With continued advances in genomic research and the increased availability of genotyping, relationships based on molecular markers are obtainable and desirable. The addition of a new program to the set of multiple-trait derivative-free REML programs is described that allows users the flexibility to calculate relationships using standard pedigree files or an arbitrary relationship matrix based on genetic marker information. The strategy behind this modification and its design is described. An application is illustrated in a QTL association study for canine hip dysplasia.

Association analysis of candidate genes for maysin and chlorogenic acid accumulation in maize silks

Two compounds, the C-glycosyl flavone maysin and the phenylpropanoid product chlorogenic acid (CGA), have been implicated in corn earworm (Helicoverpa zea Boddie) resistance in maize (Zea mays L.). Previous quantitative trait locus (QTL) analyses identified the pericarp color (p) locus, which encodes a transcription factor, as the major QTL for maysin and CGA. QTL analysis has also implicated the dihydroflavanol reductase (DFR; E.C. no. 1.1.1.219) locus anthocyaninless1 (a1) and the duplicate chalcone synthase (CHS; E.C. no. 2.3.1.74) loci colorless2 (c2) and white pollen1 (whp1) as genes underlying QTL for maysin and/or CGA synthesis. Epistatic interactions between p and a1 and between p and c2 were also defined. CHS catalyzes the first step in the flavonoid pathway and represents one of the first enzyme steps following the branch off the general phenylpropanoid pathway towards CGA synthesis. In maize, the reduction of dihydroflavanol to leucoanthocyanin by DFR immediately follows the pathway branch leading to C-glycosyl flavone production. The detection of QTLs for maysin and CGA concentration at loci encoding enzyme steps following the pathway branch points implicates alterations in the flow of biochemical intermediates as the biological basis of the QTL effects. To examine if sequence variation among alleles of a1, c2, and whp1 affect maysin and CGA synthesis in maize silks, we performed an association analysis. Because the p locus has often been a major QTL for maysin and CGA and has exhibited epistatic interactions with a1, c2, and whp1, association analysis was conditioned on the p genotype. A highly significant association of two sequence polymorphisms in the promoter of a1 with maysin synthesis was demonstrated. Additional conditioning on the genotype of the significant a1 polymorphism allowed the detection of a significant polymorphism within the whp1 promoter. Our analyses demonstrate that conditioning for epistatic factors greatly increases the power of association testing.

Structure of linkage disequilibrium and phenotypic associations in the maize genome

Association studies based on linkage disequilibrium (LD) can provide high resolution for identifying genes that may contribute to phenotypic variation. We report patterns of local and genome-wide LD in 102 maize inbred lines representing much of the worldwide genetic diversity used in maize breeding, and address its implications for association studies in maize. In a survey of six genes, we found that intragenic LD generally declined rapidly with distance (r(2) < 0.1 within 1500 bp), but rates of decline were highly variable among genes. This rapid decline probably reflects large effective population sizes in maize during its evolution and high levels of recombination within genes. A set of 47 simple sequence repeat (SSR) loci showed stronger evidence of genome-wide LD than did single-nucleotide polymorphisms (SNPs) in candidate genes. LD was greatly reduced but not eliminated by grouping lines into three empirically determined subpopulations. SSR data also supplied evidence that divergent artificial selection on flowering time may have played a role in generating population structure. Provided the effects of population structure are effectively controlled, this research suggests that association studies show great promise for identifying the genetic basis of important traits in maize with very high resolution.

Dwarf8 polymorphisms associate with variation in flowering time

Historically, association tests have been used extensively in medical genetics, but have had virtually no application in plant genetics. One obstacle to their application is the structured populations often found in crop plants, which may lead to nonfunctional, spurious associations. In this study, statistical methods to account for population structure were extended for use with quantitative variation and applied to our evaluation of maize flowering time. Mutagenesis and quantitative trait locus (QTL) studies suggested that the maize gene Dwarf8 might affect the quantitative variation of maize flowering time and plant height. The wheat orthologs of this gene contributed to the increased yields seen in the 'Green Revolution' varieties. We used association approaches to evaluate Dwarf8 sequence polymorphisms from 92 maize inbred lines. Population structure was estimated using a Bayesian analysis of 141 simple sequence repeat (SSR) loci. Our results indicate that a suite of polymorphisms associate with differences in flowering time, which include a deletion that may alter a key domain in the coding region. The distribution of nonsynonymous polymorphisms suggests that Dwarf8 has been a target of selection.

Meiotic drive of chromosomal knobs reshaped the maize genome

Meiotic drive is the subversion of meiosis so that particular genes are preferentially transmitted to the progeny. Meiotic drive generally causes the preferential segregation of small regions of the genome; however, in maize we propose that meiotic drive is responsible for the evolution of large repetitive DNA arrays on all chromosomes. A maize meiotic drive locus found on an uncommon form of chromosome 10 [abnormal 10 (Ab10)] may be largely responsible for the evolution of heterochromatic chromosomal knobs, which can confer meiotic drive potential to every maize chromosome. Simulations were used to illustrate the dynamics of this meiotic drive model and suggest knobs might be deleterious in the absence of Ab10. Chromosomal knob data from maize's wild relatives (Zea mays ssp. parviglumis and mexicana) and phylogenetic comparisons demonstrated that the evolution of knob size, frequency, and chromosomal position agreed with the meiotic drive hypothesis. Knob chromosomal position was incompatible with the hypothesis that knob repetitive DNA is neutral or slightly deleterious to the genome. We also show that environmental factors and transposition may play a role in the evolution of knobs. Because knobs occur at multiple locations on all maize chromosomes, the combined effects of meiotic drive and genetic linkage may have reshaped genetic diversity throughout the maize genome in response to the presence of Ab10. Meiotic drive may be a major force of genome evolution, allowing revolutionary changes in genome structure and diversity over short evolutionary periods.

Patterns of molecular evolution among paralogous floral homeotic genes

The plant MADS-box regulatory gene family includes several loci that control different aspects of inflorescence and floral development. Orthologs to the Arabidopsis thaliana MADS-box floral meristem genes APETALA1 and CAULIFLOWER and the floral organ identity genes APETALA3 and PISTILLATA were isolated from the congeneric species Arabidopsis lyrata. Analysis of these loci between these two Arabidopsis species, as well as three other more distantly related taxa, reveal contrasting dynamics of molecular evolution between these paralogous floral regulatory genes. Among the four loci, the CAL locus evolves at a significantly faster rate, which may be associated with the evolution of genetic redundancy between CAL and AP1. Moreover, there are significant differences in the distribution of replacement and synonymous substitutions between the functional gene domains of different floral homeotic loci. These results indicate that divergence in developmental function among paralogous members of regulatory gene families is accompanied by changes in rate and pattern of sequence evolution among loci.

Molecular evolution of type 1 serine/threonine protein phosphatases

Type 1 serine/threonine protein phosphatases (PP1s) play key roles in many cellular processes. To understand the evolutionary relationships among PP1s from various kingdoms and to provide a valid basis to evaluate the structure-function relationships of these phosphatases, 44 PP1 sequences were aligned, revealing a high sequence similarity among PP1 homologs. About one-third of the total amino acids are conserved in all the sequences studied. Most of these conserved amino acids are located within a 270-amino-acid core region. They include most sites critical to the activity and regulation of PP1s based on three-dimensional structural studies of mammalian PP1s. Positional variation analysis using a sliding window approach revealed two variable blocks in the 270-amino-acid core region. The major variable block corresponds to a subdomain composed of three alpha-helices (alphaG, alphaH, and alphaI) and three beta-sheets (beta7, beta8, and beta9). Phylogenetic analyses suggested that plant and animal PP1s form distinct monophyletic groups. The plant PP1 family contains several subgroups that may be older than the monocot-dicot divergence. In the animal PP1 family, different vertebrate isoforms appear to form distinct subgroups. Relative substitution rate studies indicated that plant PP1s are more diverse than animal PP1s, with an average substitution rate 1.5 times as large as that of animal PP1s. The possible involvement of PP1s in the establishment of multicellularity is discussed.

The evolution of ribosomal DNA: divergent paralogues and phylogenetic implications

Although nuclear ribosomal DNA (rDNA) repeats evolve together through concerted evolution, some genomes contain a considerable diversity of paralogous rDNA. This diversity includes not only multiple functional loci but also putative pseudogenes and recombinants. We examined the occurrence of divergent paralogues and recombinants in Gossypium, Nicotiana, Tripsacum, Winteraceae, and Zea ribosomal internal transcribed spacer (ITS) sequences. Some of the divergent paralogues are probably rDNA pseudogenes, since they have low predicted secondary structure stability, high substitution rates, and many deamination-driven substitutions at methylation sites. Under standard PCR conditions, the low stability paralogues amplified well, while many high-stability paralogues amplified poorly. Under highly denaturing PCR conditions (i.e., with dimethylsulfoxide), both low- and high-stability paralogues amplified well. We also found recombination between divergent paralogues. For phylogenetics, divergent ribosomal paralogues can aid in reconstructing ancestral states and thus serve as good outgroups. Divergent paralogues can also provide companion rDNA phylogenies. However, phylogeneticists must discriminate among families of divergent paralogues and recombinants or suffer from muddled and inaccurate organismal phylogenies.

Zea ribosomal repeat evolution and substitution patterns

Zea and Tripsacum nuclear ribosomal internal transcribed spacer (ITS) sequences were used to evaluate patterns of concerted evolution, rates of substitutions, patterns of methylation-induced deamination, and structural constraints of the ITS. ITS pseudogenes were identified by their phylogenetic position, differences in nucleotide composition, extensive deamination at ancestral methylation sites, and substitutions resulting in low-stability secondary RNA structures. Selection was important in shaping the kinds of polymorphisms and substitutions observed in the ITS. ITS substitution rates were significantly different among the Zea taxa. Deamination of cytosines at methylation sites was a potent mutation source, but selection appeared to maintain high methylation site density throughout the ribosomal repeat except for the gene promoter. Nucleotide divergence statistics identified selectively constrained regions at the 5' ends of the ITS1 and ITS2.

Zea systematics: ribosomal ITS evidence

Ribosomal internal transcribed spacer (ITS) sequences were used to evaluate the phylogenetics of Zea and Tripsacum. Maximum likelihood and polymorphism parsimony were used for phylogenetic reconstructions. Zea ITS nucleotide diversity was high compared to other plant species, but approximately equivalent to other maize loci. Coalescence of ITS alleles was rapid relative to other nuclear loci; however, there was still much diversity within populations. Zea and Tripsacum form a clade clearly differentiated from all other Poaceae. Four Zea ITS pseudogenes were identified by phylogenetic position and nucleotide composition. The phylogenetic position of Z. mays ssp. huehuetenangensis was clearly established as basal to the other Z. mays. The ITS phylogeny disfavored a Z. luxurians and Z. diploperennis clade, which conflicted with some previous studies. The introgression of Z. mays alleles into Z. perennis and Z. diploperennis was also established. The ITS data indicated a near contemporary divergence of domesticated maize and its two closest wild relatives.