A second unveiling: haplotig masking of the eastern oyster genome improves population-level inference

Abstract

Genome assembly can be challenging for species, such as terrestrial arthropods and broadcast-spawning marine invertebrates, that are characterized by high amounts of polymorphism and heterozygosity, large effective population sizes, and low levels of population differentiation. High levels of heterozygosity can result in genome mis-assemblies and a larger than expected genome size due to the haplotig versions of a single loci being assembled as separate loci. Here, we assembled and annotated the first chromosome-level for the eastern oyster, Crassostrea virginica. Publicly released and annotated in 2017, the assembly is still one of the most complete and contiguous molluscan genomes to date. We also developed a post hoc method to test for and mask haplotigs in published heterozygous genomes to improve the accuracy of downstream genome analysis. Masking haplotigs had a large impact on SNP discovery and estimates of nucleotide diversity and had more subtle and nuanced effects on estimates of heterozygosity, population structure analysis, and outlier detection. We show that haplotig-masking can be a powerful tool for improving genomes and we present a halpotig-masked genome assembly for the eastern oyster that will be a critical resource for foundational research into molluscan adaptation to a changing environment and for selective breeding for the aquaculture industry.

Sex‐specific gene expression in eastern oyster, Crassostrea virginica, gonad and mantle tissues

“…Reads were trimmed with the Trim Reads tool with automatic read‐through adaptor trimming enabled to remove low‐quality bases (<Q20) and fragments < 50 bp. Trimmed reads were mapped to a haplotig masked version (Puritz et al, 2023) of the Crassostrea virginica genome v3.0 (NCBI; GCA_002022765.4) using CLC Genomic Workbench's RNA‐Seq Analysis tool. To account for high polymorphism in oysters (L. Zhang, Li, et al, 2014), the length fraction setting was reduced to 0.6 as in Proestou and Sullivan (2020), while all other settings were kept at default.…”

“…Here, we leveraged a high‐quality annotated genome (Puritz et al, 2023), affordable Illumina‐based transcriptome sequencing technology, and optimized RNAseq analysis pipelines to generate global gene expression profiles for mature eastern oyster gonad tissues. Expression profiles were used to highlight sex‐specific patterns and identify conserved and novel transcripts involved in sex differentiation in the eastern oyster.…”

Understanding Crassostrea virginica tolerance of Perkinsus marinus through global gene expression analysis

“…Apart from the length fraction setting (which we reduced to 0.6 to account for high polymorphism in oysters ( Zhang L. et al, 2014 ), we used the default settings in the CLC Genomics Workbench RNA-Seq Analysis tool to map trimmed reads to the C. virginica genome v 3.0 (NCBI; GCA_002022765.4) and generate a gene expression profile for each sequenced oyster. To avoid mapping to spurious assembly artifacts, a haplotig masked version of the genome was used as the reference ( Puritz et al, 2022 ). Reads that mapped to >10 locations or in broken pairs were not counted to minimize inaccuracies in expression signals ( Katz et al, 2010 ; Anders et al, 2013 ).…”

Nucleotide and structural polymorphisms of the eastern oyster genome paint a mosaic of divergence, selection, and human impacts

“…After this first pass, three candidate intervals were expanded by relaxing the criteria to being above the 90th percentile in at least three of the six different rolling means. One detected inversion of chromosome 5 (61,560,000 to 80,150,000) intersects with a known misassembly located from approximately from 61,754,114 to 97,911,248 bp (Puritz et al 2022), so for analysis, this inversion was split into two regions, one from 61,560,000 to 64,020,000 and a second from 64,030,000 to 80,150,000 which corresponds to a linkage group that is predominantly on chromosome 6.…”

“…Briefly, reads were trimmed for low quality bases and adapter sequences using the program fastp (Chen et al 2018). Trimmed reads were then mapped to the haplotig-masked version of the eastern oyster genome (Puritz et al 2022) using bwa (Li and Durbin 2010) with modified mismatch and gap-opening parameters (-B 3 -O 5) to help compensate for the high levels of polymorphism in the oyster sequences. Duplicates were marked using Picard (Institue 2019), and subsequent BAM files were filtered with samtools (Li et al 2009) to remove low quality mappings, secondary alignments, and pcr duplicates.…”