Biology subjects

Berthel, A.

Publications and source records attributed to Berthel, A..

5 recordsLinked to original sources

bioRxiv2025

GeneCAD: Plant Genome Annotation with a DNA Foundation Model

Accurate genome annotation is fundamental to biological discovery, yet identifying gene structures directly from DNA sequence remains a major challenge in complex genomes. We introduce GeneCAD, a sequence-only framework that predicts biologically coherent gene models without requiring species-matched transcriptomic or proteomic evidence. GeneCAD integrates lineage-specific DNA representations from the PlantCAD2 foundation model with a transformer encoder and a chromosome-scale conditional random field (CRF) to enforce structural constraints, such as splice-phase and feature order. To ensure high-quality supervision, we implement a curation strategy using a sequence-based masked-motif score to filter reference transcripts. As a primary validation across diverse angiosperms, including a complex allotetraploid, GeneCAD improves transcript F1 by approximately 9% over current tools like Helixer and BRAKER3, while sharpening boundary precision and achieving a best-in-class recovery of 86% of classical coding sequences. Furthermore, we demonstrate the frameworks modularity by adapting it to animal lineages through the substitution of the underlying DNA foundation model. While the long introns of vertebrates challenge full transcript reconstruction, the model remains highly effective at identifying individual exons. By connecting evolutionary signals with structured decoding, GeneCAD provides a versatile and scalable solution for high-fidelity genome annotation across the Tree of Life. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=102 SRC="FIGDIR/small/685877v3_ufig1.gif" ALT="Figure 1"> View larger version (33K): org.highwire.dtl.DTLVardef@5dd348org.highwire.dtl.DTLVardef@1da37daorg.highwire.dtl.DTLVardef@188ef17org.highwire.dtl.DTLVardef@108da0c_HPS_FORMAT_FIGEXP M_FIG C_FIG Lay SummaryIdentifying where genes are located within a genome is a major challenge in biology, especially in plants with large and repetitive DNA. Current methods often rely on expensive laboratory data or struggle to find genes in "noisy" regions. We developed GeneCAD, a deep-learning tool that identifies genes using only the raw DNA sequence. By using AI models that recognize patterns shaped by millions of years of evolution, GeneCAD predicts gene structures with high accuracy and consistency. Our tests show that GeneCAD outperforms existing tools in plants and can be easily adapted for use in other species, including animals. By removing the need for costly lab experiments, GeneCAD provides an affordable way for researchers to quickly map the genomes of everything from rare wild plants to essential food crops.

bioinformatics↗

bioRxiv2025

PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms

Understanding how DNA sequence encodes biological function remains a fundamental challenge in biology. Flowering plants (angiosperms), the dominant terrestrial clade, exhibit maximal biochemical complexity, extraordinary species diversity (over 100,000 species), relatively recent origins ([~]160 million years), [~]200-fold variation in genome size and relative compact coding regions compared with other eukaryotes. These features present both a unique challenge and opportunity for pre-training DNA language models to understand plant-specific evolutionary conservation, regulatory architectures and genomic functions. Here, we introduce PlantCAD2, an extended context, plant-specific DNA language model with single-nucleotide resolution, pre-trained on 65 angiosperm genomes, together with a series of public benchmarks for evaluation. Comprehensive zero-shot testing shows that PlantCAD2 (676 million parameters) efficiently captures evolutionary conservation, surpassing the 7-billion-parameter Evo2 model in 10 of 12 tasks. With parameter-efficient fine-tuning, PlantCAD2 also outperforms the 1-billion-parameter AgroNT across seven cross-species tasks including chromatin accessible region, gene expression and protein translation. Moreover, its 8,192bp context window substantially improves accessible chromatin prediction in large genomes such as maize (AUPRC increasing from 0.587 to 0.711), underscoring the importance of long-range context for modeling distal regulation. Together, these results establish PlantCAD2 as a powerful, efficient, and versatile foundation model for plant genomics, enabling accurate genome annotation across diverse species.

bioinformatics↗

bioRxiv2025

Phylogenetic Analysis and Machine Learning Identify Signatures of Selection and Predict Deleterious Mutations in Common Bean

Mutations are continuous source of new alleles and genetic diversity in populations. Domestication and selection influence the accumulation of alleles occurring across a range of deleteriousness. Evidence suggests that mildly deleterious mutations (DelMut) can be purged out of breeding populations, increasing favorable allele accumulation. We used phylogeny-based analyses among 36 legume genomes to identify selection signatures and predict DelMut in common bean. We also developed a multiparent advanced generation intercrossed (MAGIC) population of black beans to characterize DelMut. Genes involved in nitrogen metabolism showed signs of positive selection in the Middle American genome, whereas genes related to phosphorylation were positively selected in the Andean genome. By combining conservation and protein information with machine learning (ML) for high-dimensional feature analysis, we characterized 82,442 sites in the MAGIC founders (36,558 polymorphic) and 4,753 sites evenly sequenced among RILs that could be potentially deleterious. Variation in the number of highly DelMut (high predicted deleterious scores) among lines was observed and later correlated with agronomic traits. Phenotypic analyses showed that calculated genetic load (and number of highly DelMut) was negatively correlated with flowering time, maturity, and yield. A detailed in-silico analysis of predicted mutations showed presence in highly conserved protein regions, which is likely to affect protein functionality. Our results show that variation in genetic load can be observed in breeding populations and potentially impact plant performance. These results contribute to understanding the genome-wide accumulation patterns of DelMut in breeding populations. Our study supports future development of strategies to reduce genetic load in promising germplasm and accelerate breeding programs. Key messages- Genome-wide highly deleterious mutations were predicted in conserved protein domains potentially affecting protein functionality. - Variation in the genetic load and number of highly deleterious mutations can be observed in artificial breeding populations. - Nitrogen- and phosphorous-related genes are under positive selection in Middle American and Andean beans respectively.

plant biology↗

bioRxiv2024

Cross-species plant genomes modeling at single nucleotide resolution using a pre-trained DNA language model

Interpreting function and fitness effects in diverse plant genomes requires transferable models. Language models (LMs) pre-trained on large-scale biological sequences can learn evolutionary conservation and offer cross-species prediction better than supervised models through fine-tuning limited labeled data. We introduce PlantCaduceus, a plant DNA LM based on the Caduceus and Mamba architectures, pre-trained on a curated dataset of 16 Angiosperm genomes. Fine-tuning PlantCaduceus on limited labeled Arabidopsis data for four tasks, including predicting translation initiation/termination sites and splice donor and acceptor sites, demonstrated high transferability to 160 million year diverged maize, outperforming the best existing DNA LM by 1.45 to 7.23-fold. PlantCaduceus is competitive to state-of-the-art protein LMs in terms of deleterious mutation identification, and is threefold better than PhyloP. Additionally, PlantCaduceus successfully identifies well-known causal variants in both Arabidopsis and maize. Overall, PlantCaduceus is a versatile DNA LM that can accelerate plant genomics and crop breeding applications.

bioinformatics↗

bioRxiv2021

Multi-tissue integrative analysis of personal epigenomes

Understanding how genetic variants impact molecular phenotypes is a key goal of functional genomics, currently hindered by reliance on a single haploid reference genome. Here, we present the EN-TEx resource of personal epigenomes, for [~]25 tissues and >10 assays in four donors (>1500 open-access functional genomic and proteomic datasets, in total). Each dataset is mapped to a matched, diploid personal genome, which has long-read phasing and structural variants. The mappings enable us to identify >1 million loci with allele-specific behavior. These loci exhibit coordinated epigenetic activity along haplotypes and less conservation than matched, non-allele-specific loci, in a fashion broadly paralleling tissue-specificity. Surprisingly, they can be accurately modelled just based on local nucleotide-sequence context. Combining EN-TEx with existing genome annotations reveals strong associations between allele-specific and GWAS loci and enables models for transferring known eQTLs to difficult-to-profile tissues. Overall, EN-TEx provides rich data and generalizable models for more accurate personal functional genomics.

genomics↗