Haplotype Assembly Research Articles

Chromosome-scale and haplotype-resolved genome assembly of Populus trichocarpa

Abstract Populus trichocarpa, a pivotal model organism for woody transgenic research, not only garners substantial scientific interest but plays an integral role in forestry economics. Previous genomic assemblies of P. trichocarpa predominantly treated its heterozygous genome as homozygous, thereby neglecting crucial haplotypic diversity. Leveraging the high fidelity (HiFi) sequencing capabilities of PacBio sequencing and the chromosome conformation capture insights provided by Illumina's Hi-C technique, this study is the first to achieve a near telomere-to-telomere assembly of both paternal and maternal haplotypes in P. trichocarpa. Comparative genomic analysis between these haplotypes has uncovered several allelic variants and pathways critical for trait determination through allele-specific expression. Furthermore, utilizing RNA-seq data from multiple tissues, this investigation has detailed the tissue-specific expression patterns of the leucine-rich repeat (LRR) gene family, which are essential in mediating plant signal transduction and developmental regulation. Our results not only illuminate the functional genomics landscape of P. trichocarpa but also provide invaluable theoretical underpinnings for the genetic improvement of woody plants and a robust framework for exploring genetic variability and allelic expression disparities in arboreal species.

Taurine pangenome uncovers a segmental duplication upstream of KIT associated with depigmentation in white-headed cattle.

Cattle have been selectively bred for coat color, spotting, and depigmentation patterns. The assumed autosomal dominant inherited genetic variants underlying the characteristic white head of Fleckvieh, Simmental, and Hereford cattle have not been identified yet, although the contribution of structural variation upstream the KIT gene has been proposed. Here, we construct a graph pangenome from 24 haplotype assemblies representing seven taurine cattle breeds to identify and characterize the white head-associated locus for the first time based on long-read sequencing data and pangenome analyses. We introduce a pangenome-wide association mapping approach which examines assembly path similarities within the graph to reveal an association between two most likely serial alleles of a complex structural variant 66 kb upstream KIT and facial depigmentation. The complex structural variant contains a variable number of tandemly duplicated 14.3 kb repeats, consisting of LTRs, LINEs, and other repetitive elements, leading to misleading alignments of short and long reads when using a linear reference. We align 250 short-read sequencing samples spanning 15 cattle breeds to the pangenome graph, further validating that the alleles of the structural variant segregate with head depigmentation. We estimate an increased count of repeats in Hereford relative to Simmental and other white-headed cattle breeds from the graph alignment coverage, suggesting a large under-assembly in the current Hereford-based cattle reference genome which had fewer copies. Our work shows that exploiting assembly path similarities within graph pangenomes can reveal trait-associated complex structural variants.

Genome assembly of wisent (Bison bonasus) uncovers a deletion that likely inactivates the THRSP gene

The wisent (Bison bonasus) is Europe’s largest land mammal. We produced a HiFi read-based wisent assembly with a contig N50 value of 91 Mb containing 99.7% of the highly conserved single copy mammalian genes which improves contiguity a thousand-fold over an existing assembly. Extended runs of homozygosity in the wisent genome compromised the separation of the HiFi reads into parental-specific read sets, which resulted in inferior haplotype assemblies. A bovine super-pangenome built with assemblies from wisent, bison, gaur, yak, taurine and indicine cattle identified a 1580 bp deletion removing the protein-coding sequence of THRSP encoding thyroid hormone-responsive protein from the wisent and bison genomes. Analysis of 725 sequenced samples across the Bovinae subfamily showed that the deletion is fixed in both Bison species but absent in Bos and Bubalus. The THRSP transcript is abundant in adipose, fat, liver, muscle, and mammary gland tissue of Bos and Bubalus, but absent in bison. This indicates that the deletion likely inactivates THRSP in bison. We show that super-pangenomes can reveal potentially trait-associated variation across phylogenies, but also demonstrate that haplotype assemblies from species that went through population bottlenecks warrant scrutiny, as they may have accumulated long runs of homozygosity that complicate phasing.

DeepHapNet: a haplotype assembly method based on RetNet and deep spectral clustering.

Gene polymorphism originates from single-nucleotide polymorphisms (SNPs), and the analysis and study of SNPs are of great significance in the field of biogenetics. The haplotype, which consists of the sequence of SNP loci, carries more genetic information than a single SNP. Haplotype assembly plays a significant role in understanding gene function, diagnosing complex diseases, and pinpointing species genes. We propose a novel method, DeepHapNet, for haplotype assembly through the clustering of reads and learning correlations between read pairs. We employ a sequence model called Retentive Network (RetNet), which utilizes a multiscale retention mechanism to extract read features and learn the global relationships among them. Based on the feature representation of reads learned from the RetNet model, the clustering process of reads is implemented using the SpectralNet model, and, finally, haplotypes are constructed based on the read clusters. Experiments with simulated and real datasets show that the method performs well in the haplotype assembly problem of diploid and polyploid based on either long or short reads. The code implementation of DeepHapNet and the processing scripts for experimental data are publicly available at https://github.com/wjj6666/DeepHapNet.

Gap-free telomere-to-telomere haplotype assembly of the tomato hind (Cephalopholis sonnerati)

The tomato hind (Cephalopholis sonnerati) is an emerging economically important grouper in recent years. With the increasing maturity of sequencing technologies and assembly methodologies, a higher quality reference genome has become both accessible and necessary. In this study, we present two telomere-to-telomere (T2T) gap-free haplotype assemblies of the tomato hind with lengths of 1039.53 Mb (YSFRI_Csonn_HA_1.0, N50 43.83 Mb) and 1039.91 Mb (YSFRI_Csonn_HB_1.0, N50 44.09 Mb). Reads from next-generation sequencing, ONT ultra-long sequencing, and PacBio HiFi sequencing exhibited mapping rates exceeding 99.8% when aligned to these two assemblies. Evaluation using Merqury indicated high accuracy for both assemblies, with average quality values of 51.80 and 51.83, respectively. Percentages of 97.9% and 97.8% of complete BUSCOs were achieved, and a total of 23,270 and 23,184 protein-code genes were inferred in each assembly. Moreover, telomere identification, centromere prediction, and repetitive sequence annotation were also successfully performed. These two assemblies provide robust foundation for the genetic analysis and development of molecular genetic breeding technologies in C. sonnerati.

Graphasing: phasing diploid genome assembly graphs with single-cell strand sequencing

Haplotype information is crucial for biomedical and population genetics research. However, current strategies to produce de novo haplotype-resolved assemblies often require either difficult-to-acquire parental data or an intermediate haplotype-collapsed assembly. Here, we present Graphasing, a workflow which synthesizes the global phase signal of Strand-seq with assembly graph topology to produce chromosome-scale de novo haplotypes for diploid genomes. Graphasing readily integrates with any assembly workflow that both outputs an assembly graph and has a haplotype assembly mode. Graphasing performs comparably to trio phasing in contiguity, phasing accuracy, and assembly quality, outperforms Hi-C in phasing accuracy, and generates human assemblies with over 18 chromosome-spanning haplotypes.

HairSplitter: haplotype assembly from long, noisy reads

HairSplitter: haplotype assembly from long, noisy reads

Strainy: phasing and assembly of strain haplotypes from long-read metagenome sequencing.

Bacterial species in microbial communities are often represented by mixtures of strains, distinguished by small variations in their genomes. Short-read approaches can be used to detect small-scale variation between strains but fail to phase these variants into contiguous haplotypes. Long-read metagenome assemblers can generate contiguous bacterial chromosomes but often suppress strain-level variation in favor of species-level consensus. Here we present Strainy, an algorithm for strain-level metagenome assembly and phasing from Nanopore and PacBio reads. Strainy takes a de novo metagenomic assembly as input and identifies strain variants, which are then phased and assembled into contiguous haplotypes. Using simulated and mock Nanopore and PacBio metagenome data, we show that Strainy assembles accurate and complete strain haplotypes, outperforming current Nanopore-based methods and comparable with PacBio-based algorithms in completeness and accuracy. We then use Strainy to assemble strain haplotypes of a complex environmental metagenome, revealing distinct strain distribution and mutational patterns in bacterial species.

The genomes of Australian wild limes

Australian wild limes occur in highly diverse range of environments and are a unique genetic resource within the genus Citrus. Here we compare the haplotype-resolved genome assemblies of six Australian native limes, including four new assemblies generated using PacBio HiFi and Hi-C sequencing data. The size of the genomes was between 315 and 391 Mb with contig N50s from 29.5 to 35 Mb. Gene completeness of the assemblies was estimated to be from 98.4 to 99.3% and the annotations from 97.7 to 98.9% based upon BUSCO, confirming the high contiguity and completeness of the assembled genomes. High collinearity was observed among the genomes and the two haplotype assemblies for each species. Gene duplication and evolutionary analysis demonstrated that the Australian citrus have undergone only one ancient whole-genome triplication event during evolution. The highest number of species-specific and expanded gene families were found in C. glauca and they were primarily enriched in purine, thiamine metabolism, amino acids and aromatic amino acids metabolism which might help C. glauca to mitigate drought, salinity, and pathogen attacks in the drier environments in which this species is found. Unique genes related to terpene biosynthesis, glutathione metabolism, and toll-like receptors in C. australasica, and starch and sucrose metabolism genes in both C. australis and C. australasica might be important candidate genes for HLB tolerance in these species. Expanded gene families were not lineage specific, however, a greater number of genes related to plant-pathogen interactions, predominantly disease resistant protein, was found in C. australasica and C. australis.

Phasing Diploid Genome Assembly Graphs with Single-Cell Strand Sequencing.

Haplotype information is crucial for biomedical and population genetics research. However, current strategies to produce de-novo haplotype-resolved assemblies often require either difficult-to-acquire parental data or an intermediate haplotype-collapsed assembly. Here, we present Graphasing, a workflow which synthesizes the global phase signal of Strand-seq with assembly graph topology to produce chromosome-scale de-novo haplotypes for diploid genomes. Graphasing readily integrates with any assembly workflow that both outputs an assembly graph and has a haplotype assembly mode. Graphasing performs comparably to trio-phasing in contiguity, phasing accuracy, and assembly quality, outperforms Hi-C in phasing accuracy, and generates human assemblies with over 18 chromosome-spanning haplotypes.

Haplotype-resolved assembly of a pig genome using single-sperm sequencing

Single gamete cell sequencing together with long-read sequencing can reliably produce chromosome-level phased genomes. In this study, we employed PacBio HiFi and Hi-C sequencing on a male Landrace pig, coupled with single-sperm sequencing of its 102 sperm cells. A haplotype assembly method was developed based on long-read sequencing and sperm-phased markers. The chromosome-level phased assembly showed higher phasing accuracy than methods that rely only on HiFi reads. The use of single-sperm sequencing data enabled the construction of a genetic map, successfully mapping the sperm motility trait to a specific region on chromosome 1 (105.40–110.70 Mb). Furthermore, with the assistance of Y chromosome-bearing sperm data, 26.16 Mb Y chromosome sequences were assembled. We report a reliable approach for assembling chromosome-level phased genomes and reveal the potential of sperm population in basic biology research and sperm phenotype research.

High-quality genome assembly enables prediction of allele-specific gene expression in hybrid poplar.

Poplar (Populus) is a well-established model system for tree genomics and molecular breeding, and hybrid poplar is widely used in forest plantations. However, distinguishing its diploid homologous chromosomes is difficult, complicating advanced functional studies on specific alleles. In this study, we applied a trio-binning design and PacBio high-fidelity long-read sequencing to obtain haplotype-phased telomere-to-telomere genome assemblies for the 2 parents of the well-studied F1 hybrid "84K" (Populus alba × Populus tremula var. glandulosa). Almost all chromosomes, including the telomeres and centromeres, were completely assembled for each haplotype subgenome apart from 2 small gaps on one chromosome. By incorporating information from these haplotype assemblies and extensive RNA-seq data, we analyzed gene expression patterns between the 2 subgenomes and alleles. Transcription bias at the subgenome level was not uncovered, but extensive-expression differences were detected between alleles. We developed machine-learning (ML) models to predict allele-specific expression (ASE) with high accuracy and identified underlying genome features most highly influencing ASE. One of our models with 15 predictor variables achieved 77% accuracy on the training set and 74% accuracy on the testing set. ML models identified gene body CHG methylation, sequence divergence, and transposon occupancy both upstream and downstream of alleles as important factors for ASE. Our haplotype-phased genome assemblies and ML strategy highlight an avenue for functional studies in Populus and provide additional tools for studying ASE and heterosis in hybrids.

A k-mer-based bulked segregant analysis approach to map seed traits in unphased heterozygous potato genomes.

High-throughput sequencing-based methods for bulked segregant analysis (BSA) allow for the rapid identification of genetic markers associated with traits of interest. BSA studies have successfully identified qualitative (binary) and quantitative trait loci (QTLs) using QTL mapping. However, most require population structures that fit the models available and a reference genome. Instead, high-throughput short-read sequencing can be combined with BSA of k-mers (BSA-k-mer) to map traits that appear refractory to standard approaches. This method can be applied to any organism and is particularly useful for species with genomes diverged from the closest sequenced genome. It is also instrumental when dealing with highly heterozygous and potentially polyploid genomes without phased haplotype assemblies and for which a single haplotype can control a trait. Finally, it is flexible in terms of population structure. Here, we apply the BSA-k-mer method for the rapid identification of candidate regions related to seed spot and seed size in diploid potato. Using a mixture of F1 and F2 individuals from a cross between 2 highly heterozygous parents, candidate sequences were identified for each trait using the BSA-k-mer approach. Using parental reads, we were able to determine the parental origin of the loci. Finally, we mapped the identified k-mers to a closely related potato genome to validate the method and determine the genomic loci underlying these sequences. The location identified for the seed spot matches with previously identified loci associated with pigmentation in potato. The loci associated with seed size are novel. Both loci are relevant in future breeding toward true seeds in potato.

Haplotype-resolved assembly of a tetraploid potato genome using long reads and low-depth offspring data

Potato is one of the world’s major staple crops, and like many important crop plants, it has a polyploid genome. Polyploid haplotype assembly poses a major computational challenge. We introduce a novel strategy for the assembly of polyploid genomes and present an assembly of the autotetraploid potato cultivar Altus. Our method uses low-depth sequencing data from an offspring population to achieve chromosomal clustering and haplotype phasing on the assembly graph. Our approach generates high-quality assemblies of individual chromosomes with haplotype-specific sequence resolution of whole chromosome arms and can be applied in common breeding scenarios where collections of offspring are available.

Trio-binning of a hinny refines the comparative organization of the horse and donkey X chromosomes and reveals novel species-specific features

We generated single haplotype assemblies from a hinny hybrid which significantly improved the gapless contiguity for horse and donkey autosomal genomes and the X chromosomes. We added over 15 Mb of missing sequence to both X chromosomes, 60 Mb to donkey autosomes and corrected numerous errors in donkey and some in horse reference genomes. We resolved functionally important X-linked repeats: the DXZ4 macrosatellite and ampliconic Equine Testis Specific Transcript Y7 (ETSTY7). We pinpointed the location of the pseudoautosomal boundaries (PAB) and determined the size of the horse (1.8 Mb) and donkey (1.88 Mb) pseudoautosomal regions (PARs). We discovered distinct differences in horse and donkey PABs: a testis-expressed gene, XKR3Y, spans horse PAB with exons1–2 located in Y and exon3 in the X–Y PAR, whereas the donkey XKR3Y is Y-specific. DXZ4 had a similar ~ 8 kb monomer in both species with 10 copies in horse and 20 in donkey. We assigned hundreds of copies of ETSTY7, a sequence horizontally transferred from Parascaris and massively amplified in equids, to horse and donkey X chromosomes and three autosomes. The findings and products contribute to molecular studies of equid biology and advance research on X-linked conditions, sex chromosome regulation and evolution in equids.

Trypanosoma cruzi STIB980: A TcI Strain for Drug Discovery and Reverse Genetics.

Since the first published genome sequence of Trypanosoma cruzi in 2005, there have been tremendous technological advances in genomics, reverse genetics, and assay development for this elusive pathogen. However, there is still an unmet need for new and better drugs to treat Chagas disease. Here, we introduce a T. cruzi assay strain that is useful for drug research and basic studies of host-pathogen interactions. T. cruzi STIB980 is a strain of discrete typing unit TcI that grows well in culture as axenic epimastigotes or intracellular amastigotes. We evaluated the optimal parameters for genetic transfection and constructed derivatives of T. cruzi STIB980 that express reporter genes for fluorescence- or bioluminescence-based drug efficacy testing, as well as a Cas9-expressing line for CRISPR/Cas9-mediated gene editing. The genome of T. cruzi STIB980 was sequenced by combining short-read Illumina with long-read Oxford Nanopore technologies. The latter served as the primary assembly and the former to correct mistakes. This resulted in a high-quality nuclear haplotype assembly of 28 Mb in 400 contigs, containing 10,043 open-reading frames with a median length of 1077 bp. We believe that T. cruzi STIB980 is a useful addition to the antichagasic toolbox and propose that it can serve as a DTU TcI reference strain for drug efficacy testing.

HLA-G 3'UTR haplotype analyses in HCV infection and HCV-derived cirrhosis, hepatocellular carcinoma and fibrosis.

Hepatitis C virus (HCV) infection is a major cause of chronic liver disease. Chronic HCV infection is also an important cause of hepatic fibrosis, cirrhosis and hepatocellular carcinoma (HCC). HCV has the capacity to evade immune surveillance by altering the host immune response. Moreover, variations in immune-related genes can lead to differential susceptibility to HCV infection as well as interfere on the susceptibility to the development of hepatic fibrosis, cirrhosis and HCC. The human leucocyte antigen G (HLA-G) gene codes for an immunomodulatory protein known to be expressed in the maternal-foetal interface and in immune-privileged tissues. The HLA-G 3' untranslated region (3'UTR) is important for mRNA stability, and variants in this region are known to impact gene expression. Studies, mainly focusing in a 14bp insertion/deletion polymorphism, have correlated HLA-G 3'UTR with susceptibility to viral infections, but other polymorphic variants in the HLA-G 3'UTR might also affect HCV infection as they are inherited as haplotypes. The present study evaluated HLA-G 3'UTR polymorphisms and performed linkage disequilibrium test and haplotype assembly in 286 HCV infected patients who have developed fibrosis, cirrhosis or HCC, as well as in 129 healthy control subjects. Haplotypes UTR-1, UTR-2 and UTR-3 were the most observed in HCV+ patients, in the frequencies of 0.276, 0.255 and 0.121, respectively. No statistically significant difference was observed between HCV+ and control subjects, even when patients were grouped according to outcome (HCC, cirrhosis or fibrosis). Despite that, some trends in the results were observed, and therefore, we cannot rule out the possibility that variants associated to high HLA-G expression can be involved in HCV infection susceptibility.

Telomere-to-telomere assemblies of 142 strains characterize the genome structural landscape in Saccharomyces cerevisiae.

Pangenomes provide access to an accurate representation of the genetic diversity of species, both in terms of sequence polymorphisms and structural variants (SVs). Here we generated the Saccharomyces cerevisiae Reference Assembly Panel (ScRAP) comprising reference-quality genomes for 142 strains representing the species' phylogenetic and ecological diversity. The ScRAP includes phased haplotype assemblies for several heterozygous diploid and polyploid isolates. We identified circa (ca.) 4,800 nonredundant SVs that provide a broad view of the genomic diversity, including the dynamics of telomere length and transposable elements. We uncovered frequent cases of complex aneuploidies where large chromosomes underwent large deletions and translocations. We found that SVs can impact gene expression near the breakpoints and substantially contribute to gene repertoire evolution. We also discovered that horizontally acquired regions insert at chromosome ends and can generate new telomeres. Overall, the ScRAP demonstrates the benefit of a pangenome in understanding genome evolution at population scale.

Pairwise comparative analysis of six haplotype assembly methods based on users’ experience

BackgroundA haplotype is a set of DNA variants inherited together from one parent or chromosome. Haplotype information is useful for studying genetic variation and disease association. Haplotype assembly (HA) is a process of obtaining haplotypes using DNA sequencing data. Currently, there are many HA methods with their own strengths and weaknesses. This study focused on comparing six HA methods or algorithms: HapCUT2, MixSIH, PEATH, WhatsHap, SDhaP, and MAtCHap using two NA12878 datasets named hg19 and hg38. The 6 HA algorithms were run on chromosome 10 of these two datasets, each with 3 filtering levels based on sequencing depth (DP1, DP15, and DP30). Their outputs were then compared.ResultRun time (CPU time) was compared to assess the efficiency of 6 HA methods. HapCUT2 was the fastest HA for 6 datasets, with run time consistently under 2 min. In addition, WhatsHap was relatively fast, and its run time was 21 min or less for all 6 datasets. The other 4 HA algorithms’ run time varied across different datasets and coverage levels. To assess their accuracy, pairwise comparisons were conducted for each pair of the six packages by generating their disagreement rates for both haplotype blocks and Single Nucleotide Variants (SNVs). The authors also compared them using switch distance (error), i.e., the number of positions where two chromosomes of a certain phase must be switched to match with the known haplotype. HapCUT2, PEATH, MixSIH, and MAtCHap generated output files with similar numbers of blocks and SNVs, and they had relatively similar performance. WhatsHap generated a much larger number of SNVs in the hg19 DP1 output, which caused it to have high disagreement percentages with other methods. However, for the hg38 data, WhatsHap had similar performance as the other 4 algorithms, except SDhaP. The comparison analysis showed that SDhaP had a much larger disagreement rate when it was compared with the other algorithms in all 6 datasets.ConclusionThe comparative analysis is important because each algorithm is different. The findings of this study provide a deeper understanding of the performance of currently available HA algorithms and useful input for other users.

Widespread allele-specific topological domains in the human genome are not confined to imprinted gene clusters

BackgroundThere is widespread interest in the three-dimensional chromatin conformation of the genome and its impact on gene expression. However, these studies frequently do not consider parent-of-origin differences, such as genomic imprinting, which result in monoallelic expression. In addition, genome-wide allele-specific chromatin conformation associations have not been extensively explored. There are few accessible bioinformatic workflows for investigating allelic conformation differences and these require pre-phased haplotypes which are not widely available.ResultsWe developed a bioinformatic pipeline, “HiCFlow,” that performs haplotype assembly and visualization of parental chromatin architecture. We benchmarked the pipeline using prototype haplotype phased Hi-C data from GM12878 cells at three disease-associated imprinted gene clusters. Using Region Capture Hi-C and Hi-C data from human cell lines (1-7HB2, IMR-90, and H1-hESCs), we can robustly identify the known stable allele-specific interactions at the IGF2-H19 locus. Other imprinted loci (DLK1 and SNRPN) are more variable and there is no “canonical imprinted 3D structure,” but we could detect allele-specific differences in A/B compartmentalization. Genome-wide, when topologically associating domains (TADs) are unbiasedly ranked according to their allele-specific contact frequencies, a set of allele-specific TADs could be defined. These occur in genomic regions of high sequence variation. In addition to imprinted genes, allele-specific TADs are also enriched for allele-specific expressed genes. We find loci that have not previously been identified as allele-specific expressed genes such as the bitter taste receptors (TAS2Rs).ConclusionsThis study highlights the widespread differences in chromatin conformation between heterozygous loci and provides a new framework for understanding allele-specific expressed genes.