Nuclear retention coupled with sequential polyadenylation dictates post-transcriptional m6A modification in the nucleus.

N6-methyladenosine (m6A) modification is deemed to be co-transcriptionally installed on pre-mRNAs, thereby influencing various downstream RNA metabolism events. However, the causal relationship between m6A modification and RNA processing is often unclear, resulting in premature or even misleading generalizations on the function of m6A modification. Here, we develop 4sU-coupled m6A-level and isoform-characterization sequencing (4sU-m6A-LAIC-seq) and 4sU-GLORI to quantify the m6A levels for both newly synthesized and steady-state RNAs at transcript and single-base-resolution levels, respectively, which enable dissecting the relationship between m6A modification and alternative RNA polyadenylation. Unexpectedly, our results show that many m6A addition events occur post-transcriptionally, especially on transcripts with high m6A levels. Importantly, we find higher m6A levels on shorter 3' UTR isoforms, which likely result from sequential polyadenylation of longer 3' UTR isoforms with prolonged nuclear dwelling time. Therefore, m6A modification can also take place post-transcriptionally to intimately couple with other key RNA metabolism processes to establish and dynamically regulate epi-transcriptomics in mammalian cells.

Integrating Iso-seq and RNA-seq data for the reannotation of the greater amberjack genome.

The greater amberjack is a very important fishery species with high commercial value, and it is distributed worldwide. Transcriptome-based studies on S. dumerili have been limited by an inadequate reference genome and a lack of well-annotated full-length transcripts. In this study, a total of 12 tissues from juvenile and adult fish both sexes were collected for next-generation RNA sequencing (RNA-seq) and full-length isoform sequencing (Iso-seq). For Iso-seq, a total of 163,218, 149,716, and 189,169 high-quality unique transcript sequences were obtained, with an N50 of 5,441, 5,255, and 5,939, from juvenile, adult male and adult female S. dumerili, respectively. We integrated the Iso-seq and RNA-seq data to construct a comprehensive gene annotation and systematically profiled the dynamics of gene expression across the 12 tissues. Our gene models had greater detail and accuracy than those from NCBI and Ensembl, with more precise polyA locations. These resources serve as a foundation for functional genomic studies and provide valuable insights into the molecular mechanisms underlying the development, reproduction and commercial traits of amberjack.

Simultaneous de novo calling and phasing of genetic variants at chromosome-scale using NanoStrand-seq.

The successful accomplishment of the first telomere-to-telomere human genome assembly, T2T-CHM13, marked a milestone in achieving completeness of the human reference genome. The upcoming era of genome study will focus on fully phased diploid genome assembly, with an emphasis on genetic differences between individual haplotypes. Most existing sequencing approaches only achieved localized haplotype phasing and relied on additional pedigree information for further whole-chromosome scale phasing. The short-read-based Strand-seq method is able to directly phase single nucleotide polymorphisms (SNPs) at whole-chromosome scale but falls short when it comes to phasing structural variations (SVs). To shed light on this issue, we developed a Nanopore sequencing platform-based Strand-seq approach, which we named NanoStrand-seq. This method allowed for de novo SNP calling with high precision (99.52%) and acheived a superior phasing accuracy (0.02% Hamming error rate) at whole-chromosome scale, a level of performance comparable to Strand-seq for haplotype phasing of the GM12878 genome. Importantly, we demonstrated that NanoStrand-seq can efficiently resolve the MHC locus, a highly polymorphic genomic region. Moreover, NanoStrand-seq enabled independent direct calling and phasing of deletions and insertions at whole-chromosome level; when applied to long genomic regions of SNP homozygosity, it outperformed the strategy that combined Strand-seq with bulk long-read sequencing. Finally, we showed that, like Strand-seq, NanoStrand-seq was also applicable to primary cultured cells. Together, here we provided a novel methodology that enabled interrogation of a full spectrum of haplotype-resolved SNPs and SVs at whole-chromosome scale, with broad applications for species with diploid or even potentially polypoid genomes.

Mapping crossover events of mouse meiotic recombination by restriction fragment ligation-based Refresh-seq.

Single-cell whole-genome sequencing methods have undergone great improvements over the past decade. However, allele dropout, which means the inability to detect both alleles simultaneously in an individual diploid cell, largely restricts the application of these methods particularly for medical applications. Here, we develop a new single-cell whole-genome sequencing method based on third-generation sequencing (TGS) platform named Refresh-seq (restriction fragment ligation-based genome amplification and TGS). It is based on restriction endonuclease cutting and ligation strategy in which two alleles in an individual cell can be cut into equal fragments and tend to be amplified simultaneously. As a new single-cell long-read genome sequencing method, Refresh-seq features much lower allele dropout rate compared with SMOOTH-seq. Furthermore, we apply Refresh-seq to 688 sperm cells and 272 female haploid cells (secondary polar bodies and parthenogenetic oocytes) from F1 hybrid mice. We acquire high-resolution genetic map of mouse meiosis recombination at low sequencing depth and reveal the sexual dimorphism in meiotic crossovers. We also phase the structure variations (deletions and insertions) in sperm cells and female haploid cells with high precision. Refresh-seq shows great performance in screening aneuploid sperm cells and oocytes due to the low allele dropout rate and has great potential for medical applications such as preimplantation genetic diagnosis.