01 PLoS Computational Biology · 17 September 2026_
PanDelos-plus: A parallel algorithm for computing sequence homology in pangenomic analysis
- Methods
- Microbial genomics
A parallel redesign of the PanDelos homology-detection algorithm cuts bacterial pangenome analysis runtime up to 14x and memory use by up to 96%, aimed at population-scale comparative genomics.
Summary
Building a pangenome starts with finding which genes across many genomes are the same gene, an all-against-all comparison that gets expensive fast as genome counts grow. PanDelos did this with an alignment-free, parameter-free k-mer approach. PanDelos-plus parallelizes the two costliest steps, Best Hit detection and Bidirectional Best Hit extraction, through data decomposition and a thread pool, and uses lighter data structures to cut memory use. On synthetic benchmarks it runs up to 14x faster and uses up to 96% less memory than the original, while returning the same gene family calls.
What to take from it
- Parallelizes Best Hit detection and Bidirectional Best Hit extraction, the two most expensive steps in k-mer based homology detection
- Up to 14x faster execution and up to 96% lower memory use on synthetic benchmarks
- Same gene family output as the original PanDelos algorithm; code at github.com/synbionics/PanDelos-plus
- Aimed at scaling k-mer homology detection to population-level bacterial pangenome datasets
Review
Benchmarks are synthetic datasets only, with no test against real, messy genome collections carrying contamination, assembly errors, or draft-quality sequences where homology calls get harder. There is no head-to-head runtime or accuracy comparison against other widely used pangenome tools such as Roary, Panaroo, or PPanGGOLiN, so it is unclear whether the speed gain matters relative to the field's actual bottlenecks. Worth checking before adopting: performance on real, large bacterial collections of thousands of genomes, and whether gene family calls stay stable across contamination and fragmented assemblies rather than just synthetic data.
02 Zoological Research · 18 September 2026_
Genomic insights into local adaptation of indigenous chickens
- Animal science
- Population genetics
Whole-genome resequencing of 1,022 chickens from 44 Chinese indigenous populations links specific climate and landcover variables to candidate adaptation genes, several of which act as regulatory elements.
Summary
The study combines ecological data on 87 agroclimatic variables with whole-genome resequencing of 1,022 chickens across 44 Chinese indigenous populations to find genomic signatures of local adaptation. Eight dominant environmental factors emerged: solar radiation, precipitation, diurnal temperature range, and five landcover variables. Landscape and comparative genomics analyses flagged known and novel candidate genes tied to specific factors, including UNC80, PTPRO, NCOR2, CSF2RB, NXT2, and PALLD, and some adaptive non-coding variants in these genes showed spatial allele-frequency gradients and acted as regulatory elements via chromatin accessibility and DNA methylation in a tissue-specific way.
What to take from it
- 1,022 chickens across 44 Chinese indigenous populations; 87 agroclimatic variables reduced to 8 dominant environmental factors
- Candidate genes linked to specific factors: UNC80 (solar radiation), PTPRO (precipitation), NCOR2 (diurnal temperature range), CSF2RB (cropland), NXT2 (tree cover), PALLD (bare ground)
- Some adaptive non-coding variants show spatial allele-frequency gradients and regulate nearby genes via chromatin accessibility and DNA methylation
- Framed as a resource for breeding climate-resilient chickens and guiding conservation strategy
Review
Gene-environment associations at this scale are correlative: landscape genomics finds candidate loci correlated with environmental gradients, not validated causal variants, and population structure that tracks geography can produce the same statistical signal without adaptation being the cause. The chromatin accessibility and methylation evidence strengthens the case for a subset of loci but is not a functional knockout or complementation test. Worth demanding before using this for breeding decisions: functional validation of the top candidate genes, and a check that the environmental associations survive stricter correction for population structure.
03 iScience · 18 September 2026_
Ancient Mongolian aurochs and cattle genomes reveal population shifts associated with the expansion of herding
- Population genetics
- Animal science
Low-coverage ancient DNA from Mongolian aurochs and early domestic cattle shows a population turnover around the arrival of herding roughly 5,000 years ago, with links to modern cattle breeds in Central China and Tibet.
Summary
Humans hunted wild aurochs (Bos primigenius) in Mongolia for millennia before domestic taurine cattle (B. taurus) arrived roughly 5,000 years ago. The authors generated low-coverage ancient genomes from Mongolian aurochs and cattle spanning this transition and compared them to cattle genomes worldwide. Individuals predating taurine cattle share ancestry with North Asian aurochs; individuals postdating the introduction instead relate to other ancient East Asian cattle, including those from the Shimao site in the Ordos region. The ancient Mongolian lineages also show distinct genetic connections to modern cattle breeds in Central China and Tibet.
What to take from it
- Ancient genomes bracket the aurochs-to-domestic-cattle transition in Mongolia around 5,000 years ago
- Pre-transition individuals share ancestry with North Asian aurochs; post-transition individuals relate to ancient East Asian cattle from sites such as Shimao in the Ordos region
- Ancient Mongolian lineages show distinct genetic connections to modern Central Chinese and Tibetan cattle breeds
- Based on low-coverage ancient DNA compared against a worldwide cattle genome panel
Review
Low-coverage ancient DNA limits resolution: population-level ancestry claims rest on methods more sensitive to reference-panel composition and sequencing depth than high-coverage genomes would be, and the abstract gives no sample size or coverage figures to judge this directly. The claimed connections to modern Central Chinese and Tibetan breeds need the underlying admixture statistics and confidence intervals, not just a qualitative description, to know how strong the signal actually is. Worth checking before treating this as settled: coverage per sample, number of individuals per time period, and whether the modern-breed connections hold up under formal admixture f-statistics rather than descriptive comparison.
04 Molecular Biology and Evolution · 16 September 2026_
Inferring the demographic history of Chinese and Indian rhesus macaque (Macaca mulatta) populations from PacBio HiFi long-read sequencing data
- Population genetics
- Methods
PacBio HiFi genomes from Chinese and Indian rhesus macaques put dates and population sizes on their split, a baseline for interpreting genetic differences between the two research-colony source populations.
Summary
Rhesus macaques used in US research colonies are sourced from either China or India, two genetically and immunologically distinct populations whose demographic history has been poorly resolved. The authors sequenced ten unrelated individuals from each population with PacBio HiFi long reads and applied four demographic inference approaches, model-based and non-model-based, checking model fit against the empirical data and accounting for mutation-rate uncertainty. They infer a split around 140,000 generations ago from an ancestral population of about 65,000 individuals, followed by divergent size histories reaching present-day estimates of about 220,000 in the Chinese population and about 14,000 in the Indian population, with gene flow between them.
What to take from it
- Ten unrelated individuals per population, PacBio HiFi long reads, four demographic inference methods cross-checked against each other
- Estimated split about 140,000 generations ago from an ancestral population of about 65,000
- Current estimated sizes: about 220,000 (China) versus about 14,000 (India), with inferred gene flow between them
- Intended as a demographic baseline for recombination mapping, selection scans, and association studies in this biomedical model species
Review
Ten individuals per population is a small sample for demographic inference, and the abstract does not say how consistent the four methods were with each other; convergent estimates would be reassuring, divergent ones would flag uncertainty invisible in a single headline number. The dependence on mutation-rate calibration is acknowledged, but any error there propagates directly into the generation-time-scaled split date and population sizes. Worth checking before using these numbers: agreement across the four inference methods, and the specific mutation rate assumed, since a change there would shift every date and size estimate in the paper.
05 Gene · 20 September 2026_
Cross-population QTL mapping and association analysis identify breeding-ready markers for high seed protein content in elite soybean accession HJ117
- Population genetics
- Methods
Two soybean RIL populations crossed with a 53%-protein accession map nine stable QTLs for seed protein content, with a follow-up GWAS narrowing one region to a marker-assisted-selection-ready SNP.
Summary
Most soybean varieties carry 35-42% seed protein; the accession HJ117 reaches 52.99% but its genetic basis was uncharacterized. Using two recombinant inbred line populations, 183 and 171 lines, crossed against HJ117 and phenotyped across three growing seasons, the authors mapped nine stable QTLs for seed protein and oil content, including three major loci on chromosomes 15 and 20 that explained 7.2-41.5% of phenotypic variance. A follow-up GWAS on 357 accessions, the two RIL populations plus parents, narrowed one region to a specific SNP and flagged three candidate genes involved in sugar transport and nutrient partitioning during seed development. Almost all (96.7%) of the protein-enhancing alleles traced back to HJ117.
What to take from it
- Two RIL populations (183 and 171 lines) phenotyped over three seasons; high heritability for protein and oil content (broad-sense h2 > 0.91)
- Nine stable QTLs, three major loci on chromosomes 15 and 20 explaining 7.2-41.5% of phenotypic variance
- GWAS on 357 accessions narrowed one locus to a single SNP and flagged candidate genes GmSWEET10, GmSop20, and POWR1
- 96.7% of protein-enhancing alleles trace to the high-protein parent HJ117
Review
Every mapping population derives from a cross with a single donor parent, HJ117, so the loci found are specific to variation present in that one accession and may not explain high protein content in other germplasm; the paper does not test transferability. The reported variance explained per locus spans a wide range, 7.2-41.5%, with LOD scores from 2.97 to 19.82, meaning some of the nine 'stable' QTLs have much weaker statistical support than others, and the abstract does not distinguish which. Worth checking before using the flagged SNP for marker-assisted selection: independent validation in a different genetic background, and the specific LOD and variance figures for the chromosome 15 locus being relied on rather than the pooled range.