01 Horticulture Research · 25 September 2026_
The heterozygous pineapple genome demonstrates the importance of haplotype-resolved plant genomes
- Structural variation
A T2T, haplotype-resolved 'Smooth Cayenne' pineapple genome finds 1,953 large structural variants between haplotypes, and shows that a 1.3 Mb inversion blocks recombination in 374 F1 seedlings while a 6 Mb one does not.
Summary
The authors assembled both haplotypes of pineapple cultivar 'Smooth Cayenne' (F180) from PacBio HiFi and Hi-C: 25 chromosomes each, 858 Mb, all centromeres and 22 of 25 telomeres resolved. The haplotypes differ by 1.5 million SNPs and 1,953 large SVs (74 inversions, 750 translocations, 1,129 segmental duplications); inversions alone cover 3-4% of the sequence. In 374 F1 seedlings, a 1.3 Mb paracentric inversion on chromosome 20 is a complete recombination coldspot, while a 6 Mb pericentric inversion on chromosome 24 still recombines at a reduced rate, probably through short double crossovers. Mapping 11,879 DArTseq markers to the phased assembly removes spurious interchromosomal linkage seen against the older collapsed assembly.
What to take from it
- Two phased haplotypes, 858 Mb, T2T for most chromosomes
- 1,953 large SVs between haplotypes; inversions reshape 3-4% of sequence
- Paracentric 1.3 Mb inversion: no recombination; pericentric 6 Mb: reduced
- Collapsed reference produced false interchromosomal linkage in marker maps
Review
The linkage-map result is the useful part for anyone running GWAS or genomic prediction on a heterozygous crop: a collapsed reference places markers wrongly and creates linkage that does not exist. The SV counts come from comparing two haplotypes of one clone, so they describe heterozygosity within one individual, not population variation; 750 translocations between homologues is high and needs checking against Hi-C contact maps and read support, since phasing switches and misjoins look the same. The double-crossover explanation for the pericentric inversion is inferred from recombination rates, not observed. Before relying on it: per-SV validation rates, whether the inversions segregate in other cultivars, and whether the cleaner linkage map actually improves prediction accuracy.
02 Nature · 23 September 2026_
Reference genomes and fossils revise bat family phylogeny and biogeography
- Population genetics
Chromosome-level long-read assemblies for 103 bat species across all 21 families, combined with 44 fossils, place bat origins in late Palaeocene Europe and reconstruct 26 ancestral bat chromosomes.
Summary
The study uses 103 chromosome-level long-read bat assemblies, 42 of them new, covering all 21 families. The resulting phylogeny places Myzopodidae as the earliest branch of Vespertilionoidea and makes Emballonuroidea and Vespertilionoidea sister groups. The authors report a mosaic history across the genome and explain why earlier phylogenies disagreed. Ancestral-state reconstruction gives 26 ancestral bat chromosomes. A 699-character morphological matrix for 65 species, including 44 pre-Quaternary fossils, was analysed together with neutral genomic sites; fossilized birth-death and biogeographic models put the origin of bats, and so of powered flight, in Europe in the late Palaeocene. Placement of the fossil Vielasia suggests laryngeal echolocation predates the crown group.
What to take from it
- 103 chromosome-level assemblies, all 21 bat families, 42 new
- Resolved Yangochiroptera relationships; Myzopodidae earliest in Vespertilionoidea
- 26 ancestral bat chromosomes
- Total-evidence dating with 44 fossils: European, late Palaeocene origin
Review
Chromosome-level assemblies across every family make this the reference set for bat comparative genomics, and the 26-chromosome ancestral karyotype is directly reusable. The phylogenetic conclusions depend on handling incomplete lineage sorting and introgression at short internal branches; "mosaic history" means gene trees disagree, so the species tree is a model output and the support for contested nodes should be checked per locus, not only as a summary. Total-evidence dating and ancestral-area inference are sensitive to fossil placement and priors; a European origin from a fossil record that is densest in Europe needs a sampling-bias test. For SV or synteny work, the assemblies are the lasting product; check per-species contiguity and annotation quality before using them.
03 Bioinformatics Advances · 26 September 2026_
A little longer, a lot better: simulation-guided exploration of extended-length single-end barcoded reads for structural variant detection
- Methods
- Structural variation
A simulator for stLFR linked reads shows that single-end barcoded reads of 1,000 bp would call SVs in HG002 nearly as well as long reads, and better than current 100 bp paired-end stLFR.
Summary
The authors wrote stLFR-sim, a Python simulator of the single-tube Long Fragment Read workflow, and used the HG002 T2T assembly to generate data under 12 configurations. They compared current paired-end 100 bp stLFR with hypothetical single-end barcoded reads of 500 and 1,000 bp. SVs were called with Aquila_stLFR and scored against the GIAB HG002 truth set with Truvari. Simulated PE100 reproduced the precision-recall pattern of real data. SV accuracy rose with read length; SE1000 was best, competitive with other linked-read and pangenome-based approaches and close to long-read methods. The simulator is MIT-licensed on GitHub.
What to take from it
- stLFR-sim: open simulator for paired-end and long single-end barcoded reads
- Benchmarked on HG002 T2T against GIAB SV truth set
- SV accuracy increases with read length; SE1000 near long-read performance
- SE500 and SE1000 are not yet sequencing products
Review
This is a design study: SE500 and SE1000 reads do not exist as a product, so the result is what a platform would gain if it could make them. Simulation matched real PE100 only in its precision-recall pattern, not necessarily in error profile, barcode collision rate or fragment-length distribution, and those drive linked-read SV calling. The benchmark is one genome and the GIAB tier-1 SV set, which excludes the hardest regions where the claimed advantage would matter. Before using it to choose a technology: per-SV-type and per-size results, performance outside GIAB confident regions, and one real SE-long dataset to calibrate the simulator.
04 MethodsX · 25 September 2026_
Resource-efficient Strand-seq library preparation in microliter volumes: a one-pot-style adaptation for standard laboratory infrastructure
- Methods
- Structural variation
A Strand-seq protocol that keeps the one-pot chemistry but runs in microliter volumes on standard 96- or 384-well plates, cutting tip use by about 60% and bead cleanups from five to two.
Summary
Strand-seq sequences template strands of single cells with their direction preserved, which gives chromosome-length haplotypes, inversion and other SV calls, and sister chromatid exchange maps. The standard microliter protocol needs five bead cleanups and per-cell handling. A previous nanoliter one-pot version removed most cleanups but needed an acoustic dispenser and custom nanoarrays. This adaptation keeps the one-pot principles (pooled MNase digestion, cumulative reagent addition, protease inactivation in place of most cleanups) in microliter volumes, compatible with manual pipetting, liquid-handling robots or acoustic dispensing.
What to take from it
- Two bead cleanups instead of five
- About 60% fewer pipette tips than the conventional protocol
- Runs in standard 96- or 384-well plates, no nanodispensing hardware
- Strand-seq is the main method for phasing and inversion calling at chromosome scale
Review
Useful for labs that want Strand-seq for inversion discovery or assembly phasing but lack the nanoliter hardware. The abstract reports consumable savings, not data quality: library complexity, fraction of usable cells, background reads and the resolution of strand-state changes against the conventional protocol are what decide whether it can replace it. It is tested on the authors' cell material; performance on primary or non-human cells, where Strand-seq needs BrdU incorporation through one division, is not shown. Ask for side-by-side cell-level QC and one SV call set compared between protocols.
05 Genes · 26 September 2026_
Whole-Genome Resequencing Reveals Genetic Diversity and Selective Sweep Signatures in Wuling Cattle
- Population genetics
- Animal science
Resequencing of Enshi, Sinan and Xiangxi cattle finds pairwise FST around 0.01 and a shared East Asian taurine-indicine ancestry, supporting one breed, with some indicine-derived regions under selection.
Summary
The authors sequenced 20 Sinan cattle and combined them with public data for 10 Enshi, 20 Xiangxi and 65 cattle from four continental reference groups, aligned to ARS-UCD1.2. The three Wuling populations share East Asian taurine and indicine ancestry and are barely differentiated (pairwise FST about 0.01), supporting classification as one breed. Single-population sweep scans and local ancestry inference found regions linked to reproduction, fat deposition, feed efficiency, skeletal and muscle development and immunity, some derived from indicine ancestry (SYTL1, TMEM222, WDTC1, SLC9A1). Cross-population scans found distinct signals between regional groups.
What to take from it
- 50 Wuling cattle plus 65 reference animals, short-read WGS
- Enshi, Sinan and Xiangxi: FST about 0.01, one breed
- Indicine-derived selected regions include SYTL1, TMEM222, WDTC1, SLC9A1
- Regional subgroups show distinct selection signals
Review
The breed-status conclusion is well supported: FST of 0.01 across three regions is clear. The selection results are weaker. Sweep scans on 10 to 20 animals per region have little power and many false positives, and the functional categories named are the usual ones for any cattle scan. Regional differences between groups of 10 to 20 may reflect sampling or drift rather than local adaptation; no environmental association test is described. Short reads against a taurine reference miss SVs and bias indicine-derived segments. Before using the candidate genes: replication in independent samples, a neutral simulation to set sweep thresholds, and a phenotype link for at least one region.