BMC Genomics最新文献

{"title":"Correction: Differential roles of coding and non-coding transcripts in obesity: insights from RNA-seq analysis of Macaca fascicularis hepatocytes.","authors":"Yu Liu, Ziting Wang, Linna Liu, Ya Tan, Wenling Zheng, Pengfei Zhang, Andrew P Hutchins, Leonard Lipovich, Weizhong Wang","doi":"10.1186/s12864-026-13170-3","DOIUrl":"10.1186/s12864-026-13170-3","url":null,"abstract":"","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13539732/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148886126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Transcriptomic analysis of eggs, rediae and cercariae reveal stage-specific adaptations in the rumen fluke Calicophoron daubneyi.","authors":"Leah R Knöpfle, Raúl O Cosentino, Markus R Schmidt, Verena K Elbert, Sandra Haug, Frank Weber, Antonio Vazquez Perera, T Nicolai Siegel, Markus Meissner","doi":"10.1186/s12864-026-13271-z","DOIUrl":"10.1186/s12864-026-13271-z","url":null,"abstract":"<p><p>Rumen flukes, particularly the trematode Calicophoron daubneyi, are emerging parasites of livestock in Europe, yet transcriptomic insights into their environmental and intermediate host stages remain limited. Here, we present a comprehensive transcriptomic analysis of eggs at three distinct developmental stages (freshly excreted, early developmental and eye-spot stages), as well as rediae and cercariae, of C. daubneyi. High-quality RNA-sequencing (RNA-seq) datasets revealed both shared and stage-specific transcriptional profiles with each developmental stage exhibiting its own distinct expression pattern. Subsequent GO-Term enrichment analyses revealed that fully embryonated eggs in eye-spot-stage especially upregulated genes related to cilia assembly, movement and motility, reflecting preparation for miracidial hatching and host-seeking behavior. Rediae showed enhanced transcription of genes involved in diverse metabolic and biosynthetic processes, supporting rapid asexual proliferation within the snail intermediate host. Cercariae exhibited predominant upregulation of genes associated with signal transduction and energy metabolism, indicating the adaptation to its changing environmental conditions. These findings provide the first transcriptomic insights into the biology of C. daubneyi outside the definitive host, reveal molecular mechanisms underlying development, transmission and adaptation to a changing environment and identify stage-specific genes as potential targets for interventions aimed at disrupting the parasites life cycle and controlling rumen flukes in the future.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13536829/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148879014","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating deep learning and pangenomics to recover missing heritability from wild structural variations.","authors":"Yiquan Wang, Minnuo Cai, Yahui Ma, Aurélien Tellier, Kai Wei","doi":"10.1186/s12864-026-13274-w","DOIUrl":"10.1186/s12864-026-13274-w","url":null,"abstract":"<p><p>Crop domestication has induced a severe genetic bottleneck that reduces the adaptive diversity present in modern cultivars. Standard intra-population genomic prediction models reliant on linear reference genomes and SNPs fail to capture the full spectrum of phenotypic variance hidden in wild relatives. Realizing this potential requires broadening the predictive paradigm from selection within narrow breeding populations toward evolutionary-scale inference across the entire wild-to-cultivated continuum. This missing heritability is largely sequestered within complex structural variations such as presence-absence and copy number variants. These variations drive environmental adaptations but remain obscured by reference bias. To recover these unmapped structural variations the field is evolving from linear coordinates to high-dimensional genomic data representations. We review this transition by contrasting explicit graph topologies that map reticulate evolution with implicit encodings like K-mers that capture sequence composition independent of alignment. Processing these complex and high-dimensional features necessitates advanced computational tools. We synthesize emerging deep learning frameworks and highlight how Graph Neural Networks resolve inheritance paths in topological data while Transformer-based foundation models extract functional syntax from sequence context. These architectures effectively integrate structural variations to resolve non-additive effects such as epistasis missed by traditional models. Computing these hidden structural variations facilitates the precise utilization of wild germplasm. We demonstrate how AI-driven strategies enable zero-shot prediction for uncharacterized wild alleles and optimize genotype-by-environment interactions. Ultimately these approaches pave the way for accelerated de novo domestication of climate-resilient crops.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-08-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13499299/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148787776","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SNP genotyping in Pseudotsuga menziesii and Pinus radiata using targeted genotyping-by-sequencing (GBS): improved Bayesian SNP calling using a beta-binomial distribution and other optimized input parameters.","authors":"Jia Guo, Gancho Slavov, Liam W Gilson, Douglas A Maguire, Anna C Magnuson, Natalie Graham, Tancred Frickey, Glenn T Howe","doi":"10.1186/s12864-026-13099-7","DOIUrl":"10.1186/s12864-026-13099-7","url":null,"abstract":"<p><strong>Background: </strong>Single-nucleotide polymorphism markers (SNPs) have important applications in gene conservation, breeding, and fundamental genetics research. Our long-term goal is to develop routine approaches for SNP genotyping in forest trees. Ideally, these approaches would be inexpensive, able to accommodate a wide range of samples and SNPs, available through commercial providers, and produce high-quality SNP data.</p><p><strong>Results: </strong>Using targeted genotyping-by-sequencing (GBS), we developed SNP assays for two highly heterozygous tree species, Douglas-fir (Pseudotsuga menziesii) and radiata pine (Pinus radiata). Using Douglas-fir haploid and diploid data, we optimized Bayesian SNP calling by testing four input parameters: (1) allele and genotype prior probabilities, (2) Rho, the beta-binomial dispersion parameter, (3) estimated read error (BayesReadError), and (4) the logPO cutoff used to filter low confidence SNP calls. logPO is the Bayesian posterior odds ratio for a called SNP. Compared to assuming a binomial distribution of read counts (Rho = 0), the beta-binomial distribution (Rho = 0.33) substantially reduced call error and heterozygote undercalling. Compared to the other Bayesian parameters, genotype priors had little effect on genotyping success. For Douglas-fir, we tested 5,360 SNP assays, and then studied the performance of the best 4,000. For radiata pine, we tested 6,000 SNP assays, and then studied the performance of the best 4,570. In Douglas-fir and radiata pine, our Bayesian approach resulted in median call rates of 95% to 98% for the top-ranked SNPs, with an estimated call error of 1.60% for known homozygous genotypes and 2.27% for known heterozygotes. In radiata pine, median and mean call rates were above 91% for GBS and SNP genotyping using an Axiom fixed genotyping array. Additionally, the median correspondence between the GBS and Axiom genotypes was about 98% overall (mean 96%).</p><p><strong>Conclusions: </strong>By optimizing Bayesian SNP calling, selecting the best 4-5 K SNPs, and excluding samples with low DNA amounts, we substantially reduced call error and heterozygote undercalling, resulting in SNP genotypes that were nearly identical to genotypes obtained using the Axiom array. Furthermore, genotyping performance should increase even further if our SNP rankings were used to develop less complex probe pools that target fewer SNPs.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13536811/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148872832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A rapid one-step NGS-based system integrating microhaplotypes and SNPs for forensic identification","authors":"Linyu Shi, Jiaxuan Hao, Jingjing Xu, Mengyang Zhao, Anjie Wei, Xinyi Zhang, Haoliang Fan, Jiangwei Yan","doi":"10.1186/s12864-026-13237-1","DOIUrl":"https://doi.org/10.1186/s12864-026-13237-1","url":null,"abstract":"Short tandem repeats (STRs) have limited performance in analysing challenging forensic samples owing to the relatively long amplicon lengths of a subset of STR loci and the presence of stutter artefacts. While single nucleotide polymorphisms (SNPs) offer advantages for degraded DNA analysis through shorter amplicons, microhaplotypes (MHs) provide higher allelic diversity per locus than diallelic SNPs, thereby improving the statistical power of mixture deconvolution. To address these needs, we developed a novel next-generation sequencing (NGS)-based MHs and SNPs system, integrating 24 MHs, 80 SNPs, and a Y-Indel marker. Additionally, we developed a rapid, one-step library preparation method, reverse-probe-guided direct amplification. This method efficiently constructs targeted NGS libraries within 2.5 h by integrating target amplification and adapter ligation in a single tube, eliminating intermediate transfers to minimise contamination risks. We conducted comprehensive forensic validation, including population genetics parameters (total discrimination power = 1–2.0265 × 10 –39 ), performance on severely degraded DNA (degradation index: 10–15, quantified by qPCR; likelihood ratios supporting suspect contribution in 1:9 mixtures with a known victim > 2.26 × 10 15 ), performance on formalin-fixed paraffin-embedded tissues (our panel achieved 100% genotyping accuracy, whereas capillary electrophoresis-STR exhibited > 38% allelic dropout), species specificity (six common non-human species: chicken, cattle, pig, mouse, rabbit and sheep, all yielded no interpretable typing data), and inhibitor resistance (evaluated against three common forensic inhibitors: humic acid, indigo and EDTA, at varying concentrations). Taken together, the developed one-step library preparation workflow and integrated MH-SNP panel offer a rapid, reliable approach for individual identification across common challenging forensic samples, including degraded DNA, degraded mixtures, FFPE tissues and inhibitor-contaminated specimens.","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148861228","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Genomic characteristics and geographical distribution of uncultivated soil prokaryotes.","authors":"Álvaro Rodríguez Del Río, Yongxing Cui, India Mansour, Matthias C Rillig","doi":"10.1186/s12864-026-13092-0","DOIUrl":"10.1186/s12864-026-13092-0","url":null,"abstract":"<p><p>Most soil prokaryotic species remain uncultivated, limiting our understanding of the terrestrial microbiome. Metagenomic sequencing, and particularly the study of metagenome-assembled genomes (MAGs), represents an unprecedented opportunity to characterize the genomic features and biogeography of uncultivated prokaryotic taxa at the large scale. Here, we analyze 40,039 genomic bins from cultivated and uncultivated soil taxa within the SMAG catalog, and examine the occurrence of uncultivated prokaryotes in 9,012 metagenomic samples from the Sandpiper resource. Compared to genera with cultivated representatives, uncultivated soil prokaryotes show smaller genomes, lower G + C content, tendency to acidophilic, non-alkaline, thermophilic and host-associated lifestyles, and slower growth rates, with the latter having the highest predictive power for cultivation status. Uncultivated soil microbes also show unique gene repertoires, characterized by a depletion of biosynthetic and motility genes. We also show that completely uncultivated genera are more abundant in tropical and arctic soils, indicating substantial hidden diversity in these regions. Our work emphasizes that current cultivation efforts systematically fail to capture a particular fraction of soil prokaryotic diversity, and provides guidelines for future cultivation strategies.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-08-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13445660/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148676965","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Phylogenetic analysis of the bacterial intracellular R-body killer proteins indicates extensive horizontal gene transfer and signature Reb sequence motifs.","authors":"Lennart Dörr, Robin Ghosh, Michael Schweikert","doi":"10.1186/s12864-026-13231-7","DOIUrl":"10.1186/s12864-026-13231-7","url":null,"abstract":"<p><p>The fascinating spiral proteinaceous structures named 'R-bodies' can be produced by a number of bacterial species and are known to cause the so-called 'killer-effect' in paramecia. The genetic determinants of the R-bodies are the 'reb genes', which are widespread among diverse proteobacteria, presumably due to horizontal gene transfer. However, the extent of their taxonomic spread, genetic sequence diversity, and gene cluster synteny has not been analyzed exhaustively using the present genetic databases. In this study we have performed an extensive genetic survey for Reb homologous proteins, including those in previously unknown taxa. Our study reveals key amino acids of Reb protein sequences that are highly conserved and may hint at the biological role of the individual Reb proteins. We also show that the genetic synteny of reb gene clusters is diverse but can be clustered into distinct groups. Further, we analyze possible horizontal gene transfer events and pathways for reb genes and indicate context with the bacterial habitat. By identifying key aspects of R-body spread and functionality with our genetic analysis we pave the way for more targeted lab experiments that will allow R-bodies to be used for biotechnological or biomedical applications.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13430693/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148667955","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparative genomics of nicotinic acetylcholine receptors reveals early divergence of the α6 subunit in cryptic Anopheles species.","authors":"Caroline Fouet, Desiree E Rios, Fred A Ashu, Matthew J Pinch, Cesar A Hernandez, Marilene M Ambadiang, Basile Kamgang, Colince Kamdem","doi":"10.1186/s12864-026-12994-3","DOIUrl":"https://doi.org/10.1186/s12864-026-12994-3","url":null,"abstract":"<p><strong>Background: </strong>Nicotinic acetylcholine receptors (nAChRs) are pentameric ligand-gated ion channels that mediate fast cholinergic transmission in the insect central nervous system and serve as targets for several widely used classes of insecticides. nAChR-targeting formulations are being deployed in malaria vector control programs to mitigate widespread resistance to pyrethroids. However, the extent of genetic variation at insecticide target sites that may facilitate evolutionary responses to chemical exposure among vector species remains poorly understood.</p><p><strong>Results: </strong>Here, we used whole-genome and Sanger sequencing to examine amino acid substitutions across all 11 nAChR subunits in sibling species of the Anopheles gambiae complex in Africa. The gene family is highly constrained, with only 33 nonsynonymous mutations detected at very low frequency in wild populations throughout the continent. No substitutions were observed at canonical ligand-binding residues within orthosteric domains, suggesting that acetylcholine and neonicotinoid binding affinities are unlikely to differ between the cryptic species. However, despite the evolutionary constraint, we identified two linked amino acid substitutions in the α6 subunit indicating early divergence of the receptor between species.</p><p><strong>Conclusions: </strong>These findings show that relaxed constraint can enable subtle amino acid changes in specific subunits of a conserved receptor family. The role of these mutations in shaping heterogeneous responses to insecticide exposure among mosquito species warrants further investigation.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13471630/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148762396","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrative population genomics and tissue-specific expression profiling in cattle using whole-genome sequence resources.","authors":"Mohammed Saeed-Zidane, Iulia Blaj, Amira Yousif, Georg Thaller","doi":"10.1186/s12864-026-13218-4","DOIUrl":"10.1186/s12864-026-13218-4","url":null,"abstract":"<p><p>Large-scale genomic resources enable systematic investigation of genetic variation and tissue-specific expression of immune-related pathways in livestock species. Here, we integrated population-level whole-genome sequence data with tissue-specific transcriptional profiling to characterize genomic differentiation and expression patterns of NF-κB pathway-related genes in cattle. Using whole-genome sequence data from the 1000 Bull Genomes Project, we identified prioritized differentiated variants within several NF-κB pathway-related genes, including CD14, NF-κB1, IL-1β, and BAFFR. The mRNA expression levels of selected NF-κB signaling-related genes were quantified across four tissues with relevance to host defense and metabolism in bulls and cows of both breeds. Expression analyses revealed pronounced tissue-dependent regulation, with distinct organ-specific transcriptional patterns across the investigated genes. Intestinal tissues showed lower expression of several innate immune genes, whereas CD14 expression was more prominent in liver-associated comparisons, highlighting functional tissue specialization within the NF-κB signaling network. Breed- and sex-associated effects were gene-dependent rather than uniform across the pathway. Histological assessment of the spleen was performed to provide an anatomical context for observed transcriptional variations. A population-prioritized CD14 missense variant (Asn177Asp), identified from the 1000 Bull Genomes dataset, was contextualized using in silico structural annotation to provide structural context for coding variation. These findings provide descriptive evidence that population-level genomic differentiation in selected NF-κB pathway genes is accompanied by tissue-specific transcriptional differences in cattle.</p>","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":"27 1","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13401305/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148590367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Systematic evaluation of single-cell foundation model interpretability: attention-derived edge scores add no incremental value over gene-level features for perturbation-target prediction.","authors":"Ihor Kendiukhov","doi":"10.1186/s12864-026-12965-8","DOIUrl":"10.1186/s12864-026-12965-8","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Single-cell foundation models such as scGPT and Geneformer are increasingly used for gene regulatory network (GRN) inference, with attention-derived edge scores routinely interpreted as regulatory proxies. Prior benchmarks have evaluated curated-reference recovery but have not systematically tested whether attention adds information beyond expression statistics for predicting the outcomes of genetic perturbations, nor whether attention-identified \"regulatory\" components are causally required for such predictions. This gap matters because the NLP interpretability literature has established that attention weights do not reliably indicate feature importance, and biological foundation models are being deployed without analogous scrutiny.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We present an evaluation framework comprising thirty-seven analyses and 153 statistical tests under Benjamini-Hochberg FDR correction, spanning two architectures (scGPT, Geneformer V2-316M), four cell types (K562, RPE1, primary T cells, iPSC neurons), and two perturbation modalities (CRISPRi, CRISPRa). The framework separates two objectives: (A) mechanistic interpretability / GRN recovery against curated references, and (B) perturbation-target prediction, i.e. classifying which genes show differential expression after a CRISPR perturbation. Five test families-trivial-baseline comparison, conditional incremental-value testing, residualisation and propensity matching, causal ablation with intervention-fidelity diagnostics, and cross-context replication-address Objective B, supplemented by a synthetic positive control establishing pipeline sensitivity.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;Attention patterns encode layer-specific biological structure-protein-protein interactions in early layers, transcriptional regulation in late layers-and Cell-State Stratified Interpretability (CSSI) exploits this structure to improve curated GRN recovery up to [Formula: see text] on the Objective A task. On Objective B, however, attention-derived edge scores add no incremental value beyond trivial gene-level features (variance, mean expression, dropout rate): gene-level baselines outperform both attention and correlation edges (AUROC 0.81-0.88 versus 0.70), augmenting gene-level predictors with pairwise edges produces ΔAUROC of [Formula: see text] to [Formula: see text] across 559,720 perturbation-gene observations, and causal ablation of TRRUST-ranked attention heads produces no degradation across three independent intervention channels. The attention-correlation relationship is context-dependent (equal in K562 CRISPRi, worse in CRISPRa, better in RPE1), but gene-level dominance on Objective B is universal across both cell types where adequate power is available.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;Attention patterns in single-cell foundation models encode biologically structured information, including layer-specific regulatory signals recoverable via CSSI, but provide no uni","PeriodicalId":9030,"journal":{"name":"BMC Genomics","volume":" ","pages":""},"PeriodicalIF":3.9,"publicationDate":"2026-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13390336/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148547753","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}