Accurate Gene Prediction Research Articles

Assessing the utility of large language models for phenotype-driven gene prioritization in the diagnosis of rare genetic disease

Phenotype-driven gene prioritization is fundamental to diagnosing rare genetic disorders. While traditional approaches rely on curated knowledge graphs with phenotype-gene relations, recent advancements in large language models (LLMs) promise a streamlined text-to-gene solution. In this study, we evaluated five LLMs, including two generative pre-trained transformers (GPT) series and three Llama2 series, assessing their performance across task completeness, gene prediction accuracy, and adherence to required output structures. We conducted experiments, exploring various combinations of models, prompts, phenotypic input types, and task difficulty levels. Our findings revealed that the best-performed LLM, GPT-4, achieved an average accuracy of 17.0% in identifying diagnosed genes within the top 50 predictions, which still falls behind traditional tools. However, accuracy increased with the model size. Consistent results were observed over time, as shown in the dataset curated after 2023. Advanced techniques such as retrieval-augmented generation (RAG) and few-shot learning did not improve the accuracy. Sophisticated prompts were more likely to enhance task completeness, especially in smaller models. Conversely, complicated prompts tended to decrease output structure compliance rate. LLMs also achieved better-than-random prediction accuracy with free-text input, though performance was slightly lower than with standardized concept input. Bias analysis showed that highly cited genes, such as BRCA1, TP53, and PTEN, are more likely to be predicted. Our study provides valuable insights into integrating LLMs with genomic analysis, contributing to the ongoing discussion on their utilization in clinical workflows.

Spaln3: improvement in speed and accuracy of genome mapping and spliced alignment of protein query sequences.

Spaln is the earliest practical tool for self-sufficient genome mapping and spliced alignment of protein query sequences onto a mammalian-sized eukaryotic genomic sequence. However, its computational speed has become inadequate for the analysis of rapidly growing genomic and transcript sequence data. The dynamic programming calculation of Spaln has been sped up in two ways: (i) the introduction of the multi-intermediate unidirectional Hirschberg method and (ii) SIMD-based vectorization. The new version, Spaln3, is ∼7 times faster than the latest Spaln version 2, and its gene prediction accuracy is consistently higher than that of Miniprot. https://github.com/ogotoh/spaln.

How to overcome constraints imposed by microsporidian genome features to ensure gene prediction?

Since the advent of sequencing techniques and due to their continuous evolution, it has become easier and less expensive to obtain the complete genome sequence of any organism. Nevertheless, to elucidate all biological processes governing organism development, quality annotation is essential. In genome annotation, predicting gene structure is one of the most important and captivating challenges for computational biology. This aspect of annotation requires continual optimization, particularly for genomes as unusual as those of microsporidia. Indeed, this group of fungal-related parasites exhibits specific features (highly reduced gene sizes, sequences with high rate of evolution) linked to their evolution as intracellular parasites, requiring the implementation of specific annotation approaches to consider all these features. This review aimed to outline these characteristics and to assess the increasingly efficient approaches and tools that have enhanced the accuracy of gene prediction for microsporidia, both in terms of sensitivity and specificity. Subsequently, a final part will be dedicated to postgenomic approaches aimed at reinforcing the annotation data generated by prediction software. These approaches include the characterization of other understudied genes, such as those encoding regulatory noncoding RNAs or very small proteins, which also play crucial roles in the life cycle of these microorganisms.

GeneMark-ETP significantly improves the accuracy of automatic annotation of large eukaryotic genomes.

Large-scale genomic initiatives, such as the Earth BioGenome Project, require efficient methods for eukaryotic genome annotation. Here we present an automatic gene finder, GeneMark-ETP, integrating genomic-, transcriptomic-, and protein-derived evidence that has been developed with a focus on large plant and animal genomes. GeneMark-ETP first identifies genomic loci where extrinsic data are sufficient for making gene predictions with "high confidence." The genes situated in the genomic space between the high-confidence genes are predicted in the next stage. The set of high-confidence genes serves as an initial training set for the statistical model. Further on, the model parameters are iteratively updated in the rounds of gene prediction and parameter re-estimation. Upon reaching convergence, GeneMark-ETP makes the final predictions and delivers the whole complement of predicted genes. GeneMark-ETP outperforms gene finders using a single type of extrinsic evidence. Comparisons with gene finders MAKER2 and TSEBRA, those that use both transcript- and protein-derived extrinsic evidence, show that GeneMark-ETP delivers state-of-the-art gene-prediction accuracy, with the margin of outperforming existing approaches increasing in its application to larger and more complex eukaryotic genomes.

Explainable Multilayer Graph Neural Network for Cancer Gene Prediction.

The identification of cancer genes is a critical yet challenging problem in cancer genomics research. Existing computational methods, including deep graph neural networks, fail to exploit the multilayered gene-gene interactions or provide limited explanations for their predictions. These methods are restricted to a single biological network, which cannot capture the full complexity of tumorigenesis. Models trained on different biological networks often yield different and even opposite cancer gene predictions, hindering their trustworthy adaptation. Here, we introduce an Explainable Multilayer Graph Neural Network (EMGNN) approach to identify cancer genes by leveraging multiple gene-gene interaction networks and pan-cancer multi-omics data. Unlike conventional graph learning on a single biological network, EMGNN uses a multilayered graph neural network to learn from multiple biological networks for accurate cancer gene prediction. Our method consistently outperforms all existing methods, with an average 7.15% improvement in area under the precision-recall curve (AUPR) over the current state-of-the-art method. Importantly, EMGNN integrated multiple graphs to prioritize newly predicted cancer genes with conflicting predictions from single biological networks. For each prediction, EMGNN provided valuable biological insights via both model-level feature importance explanations and molecular-level gene set enrichment analysis. Overall, EMGNN offers a powerful new paradigm of graph learning through modeling the multilayered topological gene relationships and provides a valuable tool for cancer genomics research. Our code is publicly available at https://github.com/zhanglab-aim/EMGNN.

Accurate and fast graph-based pangenome annotation and clustering with ggCaller.

Bacterial genomes differ in both gene content and sequence mutations, which underlie extensive phenotypic diversity, including variation in susceptibility to antimicrobials or vaccine-induced immunity. To identify and quantify important variants, all genes within a population must be predicted, functionally annotated, and clustered, representing the "pangenome." Despite the volume of genome data available, gene prediction and annotation are currently conducted in isolation on individual genomes, which is computationally inefficient and frequently inconsistent across genomes. Here, we introduce the open-source software graph-gene-caller (ggCaller). ggCaller combines gene prediction, functional annotation, and clustering into a single workflow using population-wide de Bruijn graphs, removing redundancy in gene annotation and resulting in more accurate gene predictions and orthologue clustering. We applied ggCaller to simulated and real-world bacterial data sets containing hundreds or thousands of genomes, comparing it to current state-of-the-art tools. ggCaller has considerable speed-ups with equivalent or greater accuracy, particularly with data sets containing complex sources of error, such as assembly contamination or fragmentation. ggCaller is also an important extension to bacterial genome-wide association studies, enabling querying of annotated graphs for functional analyses. We highlight this application by functionally annotating DNA sequences with significant associations to tetracycline and macrolide resistance in Streptococcus pneumoniae, identifying key resistance determinants that were missed when using only a single reference genome. ggCaller is a novel bacterial genome analysis tool with applications in bacterial evolution and epidemiology.

KOunt: a reproducible KEGG orthologue abundance workflow.

Accurate gene prediction is essential for successful metagenome analysis. We present KOunt, a Snakemake pipeline, that precisely quantifies KEGG orthologue abundance. KOunt is available on GitHub: https://github.com/WatsonLab/KOunt. The KOunt reference database is available on figshare: https://doi.org/10.6084/m9.figshare.21269715. Test data are available at https://doi.org/10.6084/m9.figshare.22250152 and version 1.2.0 of KOunt at https://doi.org/10.6084/m9.figshare.23607834.

Eye2Gene

AbstractPurpose: Inherited retinal diseases (IRDs) are single‐gene disorders caused by genetic mutations in any one of over 270 genes. Identifying the causative gene through genetic testing is crucial for gene targeted treatments, recruitment to clinical trials, prognosis and family planning. The prescription and interpretation of genetic results requires phenotype–genotype recognition that only IRD experts can provide hence this has motivated AI approaches that are able to predict the probable IRD causative gene from the retinal scans of suspected IRD patients. However, these AI approaches are currently “black boxes” that do not offer any clinical interpretability nor do they provide fine‐grained phenotypic information which is essential for prognosis. We therefore sought to develop an AI algorithm capable of automatically identifying and quantifying IRD‐specific features in retinal scans.Methods: In order to build a training dataset for the AI algorithm, a grading protocol was drafted defining retinal features important in the identification of the 36 most common IRD genes. Optical Coherence Tomography (OCT) and Fundus Autofluorescence (FAF) scans were manually segmented by four graders over three rounds of grading including adjudication, feedback and protocol clarification where required. This iterative process was followed to assure good agreement, assessed using the dice score metric, between features. Features that were too difficult or laborious to annotate were converted to labels. Using the manually segmented data, an AI algorithm known as a U‐net, was trained to automatically segment 15 features. The number, size and brightness of the automatically identified features was quantified and compared across the 36 gene classes.Results: A total of 3527 scan‐features were manually annotated across scans of 36 genes. The inter‐grader dice scores ranged from 0.30 to 0.91 with an average 0.54. Features with the best agreement were anatomical features such as whole retina on OCT (0.91) and optic disc on FAF (0.87), and yielded the best predictions by the segmentation model 0.87 and 0.82, respectively. Pathological features with good inter‐grader agreement included ellipsoid zone loss (0.78) and hypo autofluorescence (0.65) and those with poorer inter‐grader agreement included hyper autofluorescence (0.30) and retinal pigment epithelium loss (0.42). The segmentation model achieved an average dice score of 0.71 across all features. Statistically significant differences were found in feature count, size and brightness between the 36 different gene classes and prediction accuracy of 58% was achieved using random forests to predict the correct causative gene from 36 genes using these features. Using a “black box” AI approach the gene prediction accuracy was 66%.Conclusions: Automated segmentation of features in IRD scans using AI is feasible and leads to interpretable prediction of disease‐associated IRD genes. However, “black box” prediction can still achieve higher accuracy at the expense of interpretability. Further work is needed in identifying and standardizing the definition of IRD retinal features to improve automatic segmentation accuracy and the AI‐assisted prediction of IRD‐associated genes.

Complete Genome Sequence Resource for Xanthomonas translucens pv. undulosa MAI5034, a Wheat Pathogen from Uruguay.

Complete Genome Sequence Resource for Xanthomonas translucens pv. undulosa MAI5034, a Wheat Pathogen from Uruguay.

Machine learning for identifying resistance features of Klebsiella pneumoniae using whole-genome sequence single nucleotide polymorphisms.

Introduction. Klebsiella pneumoniae, a gram-negative bacterium, is a common pathogen causing nosocomial infection. The drug-resistance rate of K. pneumoniae is increasing year by year, posing a severe threat to public health worldwide. K. pneumoniae has been listed as one of the pathogens causing the global crisis of antimicrobial resistance in nosocomial infections. We need to explore the drug resistance of K. pneumoniae for clinical diagnosis. Single nucleotide polymorphisms (SNPs) are of high density and have rich genetic information in whole-genome sequencing (WGS), which can affect the structure or expression of proteins. SNPs can be used to explore mutation sites associated with bacterial resistance.Hypothesis/Gap Statement. Machine learning methods can detect genetic features associated with the drug resistance of K. pneumoniae from whole-genome SNP data.Aims. This work used Fast Feature Selection (FFS) and Codon Mutation Detection (CMD) machine learning methods to detect genetic features related to drug resistance of K. pneumoniae from whole-genome SNP data.Methods. WGS data on resistance of K. pneumoniae strains to four antibiotics (tetracycline, gentamicin, imipenem, amikacin) were downloaded from the European Nucleotide Archive (ENA). Sequence alignments were performed with MUMmer 3 to complete SNP calling using K. pneumoniae HS11286 chromosome as the reference genome. The FFS algorithm was applied to feature selection of the SNP dataset. The training set was constructed based on mutation sites with mutation frequency >0.995. Based on the original SNP training set, 70% of SNPs were randomly selected from each dataset as the test set to verify the accuracy of the training results. Finally, the resistance genes were obtained by the CMD algorithm and Venny.Results. The number of strains resistant to tetracycline, gentamicin, imipenem and amikacin was 931, 1048, 789 and 203, respectively. Machine learning algorithms were applied to the SNP training set and test set, and 28 and 23 resistance genes were predicted, respectively. The 28 resistance genes in the training set included 22 genes in the test set, which verified the accuracy of gene prediction. Among them, some genes (KPHS_35310, KPHS_18220, KPHS_35880, etc.) corresponded to known resistance genes (Eef2, lpxK, MdtC, etc). Logistic regression classifiers were established based on the identified SNPs in the training set. The area under the curves (AUCs) of the four antibiotics was 0.939, 0.950, 0.912 and 0.935, showing a strong ability to predict bacterial resistance.Conclusion. Machine learning methods can effectively be used to predict resistance genes and associated SNPs. The FFS and CMD algorithms have wide applicability. They can be used for the drug-resistance analysis of any microorganism with genomic variation and phenotypic data. This work lays a foundation for resistance research in clinical applications.

Improved supervised prediction of aging-related genes via weighted dynamic network analysis

BackgroundThis study focuses on the task of supervised prediction of aging-related genes from -omics data. Unlike gene expression methods for this task that capture aging-specific information but ignore interactions between genes (i.e., their protein products), or protein–protein interaction (PPI) network methods for this task that account for PPIs but the PPIs are context-unspecific, we recently integrated the two data types into an aging-specific PPI subnetwork, which yielded more accurate aging-related gene predictions. However, a dynamic aging-specific subnetwork did not improve prediction performance compared to a static aging-specific subnetwork, despite the aging process being dynamic. This could be because the dynamic subnetwork was inferred using a naive Induced subgraph approach. Instead, we recently inferred a dynamic aging-specific subnetwork using a methodologically more advanced notion of network propagation (NP), which improved upon Induced dynamic aging-specific subnetwork in a different task, that of unsupervised analyses of the aging process.ResultsHere, we evaluate whether our existing NP-based dynamic subnetwork will improve upon the dynamic as well as static subnetwork constructed by the Induced approach in the considered task of supervised prediction of aging-related genes. The existing NP-based subnetwork is unweighted, i.e., it gives equal importance to each of the aging-specific PPIs. Because accounting for aging-specific edge weights might be important, we additionally propose a weighted NP-based dynamic aging-specific subnetwork. We demonstrate that a predictive machine learning model trained and tested on the weighted subnetwork yields higher accuracy when predicting aging-related genes than predictive models run on the existing unweighted dynamic or static subnetworks, regardless of whether the existing subnetworks were inferred using NP or the Induced approach.ConclusionsOur proposed weighted dynamic aging-specific subnetwork and its corresponding predictive model could guide with higher confidence than the existing data and models the discovery of novel aging-related gene candidates for future wet lab validation.

Relationship between gene regulation network structure and prediction accuracy in high dimensional regression

The least absolute shrinkage and selection operator (lasso) and principal component regression (PCR) are popular methods of estimating traits from high-dimensional omics data, such as transcriptomes. The prediction accuracy of these estimation methods is highly dependent on the covariance structure, which is characterized by gene regulation networks. However, the manner in which the structure of a gene regulation network together with the sample size affects prediction accuracy has not yet been sufficiently investigated. In this study, Monte Carlo simulations are conducted to investigate the prediction accuracy for several network structures under various sample sizes. When the gene regulation network is a random graph, a sufficiently large number of observations are required to ensure good prediction accuracy with the lasso. The PCR provided poor prediction accuracy regardless of the sample size. However, a real gene regulation network is likely to exhibit a scale-free structure. In such cases, the simulation indicates that a relatively small number of observations, such as N=300, is sufficient to allow the accurate prediction of traits from a transcriptome with the lasso.

Supervised Prediction of Aging-Related Genes From a Context-Specific Protein Interaction Subnetwork.

Human aging is linked to many prevalent diseases. The aging process is highly influenced by genetic factors. Hence, it is important to identify human aging-related genes. We focus on supervised prediction of such genes. Gene expression-based methods for this purpose study genes in isolation from each other. While protein-protein interaction (PPI) network-based methods for this purpose account for interactions between genes' protein products, current PPI network data are context-unspecific, spanning different biological conditions. Instead, here, we focus on an aging-specific subnetwork of the entire PPI network, obtained by integrating aging-specific gene expression data and PPI network data. The potential of such data integration has been recognized but mostly in the context of cancer. So, we are the first to propose a supervised learning framework for predicting aging-related genes from an aging-specific PPI subnetwork. In a systematic and comprehensive evaluation, we find that in many of the evaluation tests: (i) using an aging-specific subnetwork indeed yields more accurate aging-related gene predictions than using the entire network, and (ii) predictive methods from our framework that have not previously been used for supervised prediction of aging-related genes outperform existing prominent methods for the same purpose. These results justify the need for our framework.

GeneMark-HM: improving gene prediction in DNA sequences of human microbiome.

Computational reconstruction of nearly complete genomes from metagenomic reads may identify thousands of new uncultured candidate bacterial species. We have shown that reconstructed prokaryotic genomes along with genomes of sequenced microbial isolates can be used to support more accurate gene prediction in novel metagenomic sequences. We have proposed an approach that used three types of gene prediction algorithms and found for all contigs in a metagenome nearly optimal models of protein-coding regions either in libraries of pre-computed models or constructed de novo. The model selection process and gene annotation were done by the new GeneMark-HM pipeline. We have created a database of the species level pan-genomes for the human microbiome. To create a library of models representing each pan-genome we used a self-training algorithm GeneMarkS-2. Genes initially predicted in each contig served as queries for a fast similarity search through the pan-genome database. The best matches led to selection of the model for gene prediction. Contigs not assigned to pan-genomes were analyzed by crude, but still accurate models designed for sequences with particular GC compositions. Tests of GeneMark-HM on simulated metagenomes demonstrated improvement in gene annotation of human metagenomic sequences in comparison with the current state-of-the-art gene prediction tools.

VirION2: a short- and long-read sequencing and informatics workflow to study the genomic diversity of viruses in nature.

Microbes play fundamental roles in shaping natural ecosystem properties and functions, but do so under constraints imposed by their viral predators. However, studying viruses in nature can be challenging due to low biomass and the lack of universal gene markers. Though metagenomic short-read sequencing has greatly improved our virus ecology toolkit—and revealed many critical ecosystem roles for viruses—microdiverse populations and fine-scale genomic traits are missed. Some of these microdiverse populations are abundant and the missed regions may be of interest for identifying selection pressures that underpin evolutionary constraints associated with hosts and environments. Though long-read sequencing promises complete virus genomes on single reads, it currently suffers from high DNA requirements and sequencing errors that limit accurate gene prediction. Here we introduce VirION2, an integrated short- and long-read metagenomic wet-lab and informatics pipeline that updates our previous method (VirION) to further enhance the utility of long-read viral metagenomics. Using a viral mock community, we first optimized laboratory protocols (polymerase choice, DNA shearing size, PCR cycling) to enable 76% longer reads (now median length of 6,965 bp) from 100-fold less input DNA (now 1 nanogram). Using a virome from a natural seawater sample, we compared viromes generated with VirION2 against other library preparation options (unamplified, original VirION, and short-read), and optimized downstream informatics for improved long-read error correction and assembly. VirION2 assemblies combined with short-read based data (‘enhanced’ viromes), provided significant improvements over VirION libraries in the recovery of longer and more complete viral genomes, and our optimized error-correction strategy using long- and short-read data achieved 99.97% accuracy. In the seawater virome, VirION2 assemblies captured 5,161 viral populations (including all of the virus populations observed in the other assemblies), 30% of which were uniquely assembled through inclusion of long-reads, and 22% of the top 10% most abundant virus populations derived from assembly of long-reads. Viral populations unique to VirION2 assemblies had significantly higher microdiversity means, which may explain why short-read virome approaches failed to capture them. These findings suggest the VirION2 sample prep and workflow can help researchers better investigate the virosphere, even from challenging low-biomass samples. Our new protocols are available to the research community on protocols.io as a ‘living document’ to facilitate dissemination of updates to keep pace with the rapid evolution of long-read sequencing technology.

InsectOR-Webserver for sensitive identification of insect olfactory receptor genes from non-model genomes.

Insect Olfactory Receptors (ORs) are diverse family of membrane protein receptors responsible for most of the insect olfactory perception and communication, and hence they are of utmost importance for developing repellents or pesticides. Accurate gene prediction of insect ORs from newly sequenced genomes is an important but challenging task. We have developed a dedicated webserver, 'insectOR', to predict and validate insect OR genes using multiple gene prediction algorithms, accompanied by relevant validations. It is possible to employ this server nearly automatically and perform rapid prediction of the OR gene loci from thousands of OR-protein-to-genome alignments, resolve gene boundaries for tandem OR genes and refine them further to provide more complete OR gene models. InsectOR outperformed the popular genome annotation pipelines (MAKER and NCBI eukaryotic genome annotation) in terms of overall sensitivity at base, exon and locus level, when tested on two distantly related insect genomes. It displayed more than 95% nucleotide level precision in both tests. Finally, given the same input data and parameters, InsectOR missed less than 2% gene loci, in contrast to 55% loci missed by MAKER for Drosophila melanogaster. The webserver is freely available on the web at http://caps.ncbs.res.in/insectOR/ and the basic package can be downloaded from https://github.com/sdk15/insectOR for local use. This tool will allow biologists to perform quick preliminary identification of insect olfactory receptor genes from newly sequenced genomes and also assist in their further detailed annotation. Its usage can also be extended to other divergent gene families.

Complete representation of a tapeworm genome reveals chromosomes capped by centromeres, necessitating a dual role in segregation and protection

BackgroundChromosome-level assemblies are indispensable for accurate gene prediction, synteny assessment, and understanding higher-order genome architecture. Reference and draft genomes of key helminth species have been published, but little is yet known about the biology of their chromosomes. Here, we present the complete genome of the tapeworm Hymenolepis microstoma, providing a reference quality, end-to-end assembly that represents the first fully assembled genome of a spiralian/lophotrochozoan, revealing new insights into chromosome evolution.ResultsLong-read sequencing and optical mapping data were added to previous short-read data enabling complete re-assembly into six chromosomes, consistent with karyology. Small genome size (169 Mb) and lack of haploid variation (1 SNP/3.2 Mb) contributed to exceptionally high contiguity with only 85 gaps remaining in regions of low complexity sequence. Resolution of repeat regions reveals novel gene expansions, micro-exon genes, and spliced leader trans-splicing, and illuminates the landscape of transposable elements, explaining observed length differences in sister chromatids. Syntenic comparison with other parasitic flatworms shows conserved ancestral linkage groups indicating that the H. microstoma karyotype evolved through fusion events. Strikingly, the assembly reveals that the chromosomes terminate in centromeric arrays, indicating that these motifs play a role not only in segregation, but also in protecting the linear integrity and full lengths of chromosomes.ConclusionsDespite strong conservation of canonical telomeres, our results show that they can be substituted by more complex, species-specific sequences, as represented by centromeres. The assembly provides a robust platform for investigations that require complete genome representation.

Integration of anatomy ontology data with protein\u2013protein interaction networks improves the candidate gene prediction accuracy for anatomical entities

BackgroundIdentification of genes responsible for anatomical entities is a major requirement in many fields including developmental biology, medicine, and agriculture. Current wet lab techniques used for this purpose, such as gene knockout, are high in resource and time consumption. Protein–protein interaction (PPI) networks are frequently used to predict disease genes for humans and gene candidates for molecular functions, but they are rarely used to predict genes for anatomical entities. Moreover, PPI networks suffer from network quality issues, which can be a limitation for their usage in predicting candidate genes. Therefore, we developed an integrative framework to improve the candidate gene prediction accuracy for anatomical entities by combining existing experimental knowledge about gene-anatomical entity relationships with PPI networks using anatomy ontology annotations. We hypothesized that this integration improves the quality of the PPI networks by reducing the number of false positive and false negative interactions and is better optimized to predict candidate genes for anatomical entities. We used existing Uberon anatomical entity annotations for zebrafish and mouse genes to construct gene networks by calculating semantic similarity between the genes. These anatomy-based gene networks were semantic networks, as they were constructed based on the anatomy ontology annotations that were obtained from the experimental data in the literature. We integrated these anatomy-based gene networks with mouse and zebrafish PPI networks retrieved from the STRING database and compared the performance of their network-based candidate gene predictions.ResultsAccording to evaluations of candidate gene prediction performance tested under four different semantic similarity calculation methods (Lin, Resnik, Schlicker, and Wang), the integrated networks, which were semantically improved PPI networks, showed better performances by having higher area under the curve values for receiver operating characteristic and precision-recall curves than PPI networks for both zebrafish and mouse.ConclusionIntegration of existing experimental knowledge about gene-anatomical entity relationships with PPI networks via anatomy ontology improved the candidate gene prediction accuracy and optimized them for predicting candidate genes for anatomical entities.

"Mind the Gap": Hi-C Technology Boosts Contiguity of the Globe Artichoke Genome in Low-Recombination Regions.

Globe artichoke (Cynara cardunculus var. scolymus; 2n2x=34) is cropped largely in the Mediterranean region, being Italy the leading world producer; however, over time, its cultivation has spread to the Americas and China. In 2016, we released the first (v1.0) globe artichoke genome sequence (http://www.artichokegenome.unito.it/). Its assembly was generated using ∼133-fold Illumina sequencing data, covering 725 of the 1,084 Mb genome, of which 526 Mb (73%) were anchored to 17 chromosomal pseudomolecules. Based on v1.0 sequencing data, we generated a new genome assembly (v2.0), obtained from a Hi-C (Dovetail) genomic library, and which improves the scaffold N50 from 126 kb to 44.8 Mb (∼356-fold increase) and N90 from 29 kb to 17.8 Mb (∼685-fold increase). While the L90 of the v1.0 sequence included 6,123 scaffolds, the new v2.0 just 15 super-scaffolds, a number close to the haploid chromosome number of the species. The newly generated super-scaffolds were assigned to pseudomolecules using reciprocal blast procedures. The cumulative size of unplaced scaffolds in v2.0 was reduced of 165 Mb, increasing to 94% the anchored genome sequence. The marked improvement is mainly attributable to the ability of the proximity ligation-based approach to deal with both heterochromatic (e.g.: peri-centromeric) and euchromatic regions during the assembly procedure, which allowed to physically locate low recombination regions. The new high-quality reference genome enhances the taxonomic breadth of the data available for comparative plant genomics and led to a new accurate gene prediction (28,632 genes), thus promoting the map-based cloning of economically important genes.

GeneMark-EP+: eukaryotic gene prediction with self-training in the space of genes and proteins.

We have made several steps toward creating a fast and accurate algorithm for gene prediction in eukaryotic genomes. First, we introduced an automated method for efficient ab initio gene finding, GeneMark-ES, with parameters trained in iterative unsupervised mode. Next, in GeneMark-ET we proposed a method of integration of unsupervised training with information on intron positions revealed by mapping short RNA reads. Now we describe GeneMark-EP, a tool that utilizes another source of external information, a protein database, readily available prior to the start of a sequencing project. A new specialized pipeline, ProtHint, initiates massive protein mapping to genome and extracts hints to splice sites and translation start and stop sites of potential genes. GeneMark-EP uses the hints to improve estimation of model parameters as well as to adjust coordinates of predicted genes if they disagree with the most reliable hints (the -EP+ mode). Tests of GeneMark-EP and -EP+ demonstrated improvements in gene prediction accuracy in comparison with GeneMark-ES, while the GeneMark-EP+ showed higher accuracy than GeneMark-ET. We have observed that the most pronounced improvements in gene prediction accuracy happened in large eukaryotic genomes.