Bioinformatics Advances

Abstract Motivation Although machine learning models are commonly used in medical research, many analyses implement a simple partition into training data and hold-out test data, with cross-validation (CV) for tuning of model hyperparameters. Nested CV with embedded feature selection is especially suited to biomedical data where the sample size is frequently limited, but the number of predictors may be significantly larger (P ≫ n). Results The nestedcv R package implements fully nested k × l-fold CV for lasso and elastic-net regularized linear models via the glmnet package and supports a large array of other machine learning models via the caret framework. Inner CV is used to tune models and outer CV is used to determine model performance without bias. Fast filter functions for feature selection are provided and the package ensures that filters are nested within the outer CV loop to avoid information leakage from performance test sets. Measurement of performance by outer CV is also used to implement Bayesian linear and logistic regression models using the horseshoe prior over parameters to encourage a sparse model and determine unbiased model accuracy. Availability and implementation The R package nestedcv is available from CRAN: https://CRAN.R-project.org/package=nestedcv.

Myles J Lewis

Athina Spiliopoulou

Katriona Goldmann

Costantino Pitzalis

Paul McKeigue

Michael R Barnes

nestedcv: an R package for fast implementation of nested cross-validation with embedded feature selection designed for transcriptomics and high-dimensional data

nestedcv：一种用于转录组学和高维数据嵌入特征选择快速实现嵌套交叉验证的R包

Abstract Motivation Direct RNA-seq (dRNA-seq) using Oxford Nanopore Technology (ONT) has revolutionized transcript mapping by offering enhanced precision due to its long-read length. Unlike traditional techniques, dRNA-seq eliminates the need for PCR amplification, reducing the impact of GC bias, and preserving valuable base physical information, such as RNA modification and poly(A) length estimation. However, the rapid advancement of ONT devices has set higher standards for analytical software, resulting in potential challenges of software incompatibility and reduced efficiency. Results We present a novel workflow, called FASTdRNA, to manipulate dRNA-seq data efficiently. This workflow comprises two modules: a data preprocessing module and a data analysis module. The preprocessing data module, dRNAmain, encompasses basecalling, mapping, and transcript counting, which are essential for subsequent analyses. The data analysis module consists of a range of downstream analyses that facilitate the estimation of poly(A) length, prediction of RNA modifications, and assessment of alternative splicing events across different conditions with duplication. The FASTdRNA workflow is designed for the Snakemake framework and can be efficiently executed locally or in the cloud. Comparative experiments have demonstrated its superior performance compared to previous methods. This innovative workflow enhances the research capabilities of dRNA-seq data analysis pipelines by optimizing existing processes and expanding the scope of analysis. Availability and implementation The workflow is freely available at https://github.com/Tomcxf/FASTdRNA under an MIT license. Detailed install and usage guidance can be found in the GitHub repository.

Xiaofeng Chen

Yongqi Liu

Kaiwen Lv

Meiling Wang

Xiaoqin Liu

Bosheng Li

FASTdRNA: a workflow for the analysis of ONT direct RNA sequencing

FASTdRNA：一种用于ONT直接RNA测序分析的工作流程

Abstract Motivation This study investigates the flexible refinement of AlphaFold2 models against corresponding cryo-electron microscopy (cryo-EM) maps using normal modes derived from elastic network models (ENMs) as basis functions for displacement. AlphaFold2 generally predicts highly accurate structures, but 18 of the 137 models of isolated chains exhibit a TM-score below 0.80. We achieved a significant improvement in four of these deviating structures and used them to systematically optimize the parameters of the ENM motion model. Results We successfully refined four AlphaFold2 models with notable discrepancies: lipid-preserved respiratory supercomplex (TM-score increased from 0.52 to 0.69), flagellar L-ring protein (TM-score increased from 0.53 to 0.64), cation diffusion facilitator YiiP (TM-score increased from 0.76 to 0.83), and Sulfolobus islandicus pilus (TM-score increased from 0.77 to 0.85). We explored the effect of three different mode ranges (modes 1–9, 7–9, and 1–12), masked or box-cropped density maps, numerical optimization methods, and two similarity measures (Pearson correlation and inner product). The best results were achieved for the widest mode range (modes 1–12), masked maps, inner product, and local Powell optimization. These optimal parameters were implemented in the flexible fitting utility elforge.py in version 1.4 of our Python-based ModeHunter package. Availability https://modehunter.biomachina.org.

Maytha Alshammari

Jing He

Willy Wriggers

Flexible fitting of AlphaFold2-Predicted models to Cryo-EM density maps using elastic network models: a methodical affirmation

使用弹性网络模型将AlphaFold2预测的模型灵活拟合到冷冻电镜密度图：一种方法论的肯定

Abstract Motivation Developing competency in the broad area of bioinformatics is challenging globally, owing to the breadth of the field and the diversity of its audiences for education and training. Course design can be facilitated by the use of a competency framework—a set of competency requirements that define the knowledge, skills and attitudes needed by individuals in (or aspiring to be in) a particular profession or role. These competency requirements can help to define curricula as they can inform both the content and level to which competency needs to be developed. The International Society for Computational Biology (ISCB) developed a list of bioinformatics competencies in 2014, and these have undergone several rounds of improvement. In consultation with a broad bioinformatics training community, these have now been further refined and extended to include knowledge skills and attitudes, and mappings to previous and other existing competency frameworks. Results Here we present version 3 of the ISCB competency framework. We describe how it was developed and how to access it, as well as providing some examples of how it has been used. Availability The framework is openly accessible at https://competency.ebi.ac.uk/framework/iscb/3.0/competencies Supplementary information Supplementary information is available as follows: 1. Mapping version history 2. Competency guidelines DOI: 10.5281/zenodo.5418103 3. Use cases DOI: 10.5281/zenodo.5810303

Cath Brooksbank

Michelle D Brazas

Nicola Mulder

Russell Schwartz

Verena Ras

Sarah L Morgan

Marta Lloret-Llinares

Patricia Carvajal-López

Lee Larcombe

Amel Ghouila

Tom Hancocks

Venkata Satagopam

Javier De Las Rivas

Gaston Mazandu

Bruno Gaeta

The ISCB competency framework v. 3: a revised and extended standard for bioinformatics education and training

ISCB能力框架v.3：生物信息学教育和培训的修订与扩展标准

Abstract Summary Transposable Elements (TEs), commonly referred to as 'mobile elements,' constitute DNA segments capable of relocating within a genome. Initially disregarded as 'junk DNA' devoid of specific functionality, it has become evident that TEs have diverse influences on an organism's biology and health. The impact of these elements varies according to their location, classification, and their effects on specific genes or regulatory components. Despite their significant roles, a paucity of resources concerning TEs in population-scale genome sequencing remains. Herein, we analyze whole-genome sequencing data sourced from the Korean Genome and Epidemiology Study, encompassing 2,500 Korean individuals. To facilitate convenient data access and observation, we developed a web-based database, KTED. Additionally, we scrutinized the differential distributions of TEs across five distinct common disease groups: dyslipidemia, hypertension, diabetes, thyroid disease, and cancer. Availability https://snubh.shinyapps.io/KTED Supplementary information Supplementary data are available at Bioinformatics online.

Jin-Ok Lee

Sejoon Lee

Dongyoon Lee

Taeyeon Hwang

Soobok Joe

Jin Ok Yang

Jibin Jeong

Jung Hun Ohn

Jee Hyun Kim

KTED: a comprehensive Web-Based database for transposable elements in the korean genome

KTED：一个综合性的基于网络的数据库，用于韩国基因组中的转座元件

Abstract Motivation Gene expression prediction plays a vital role in transcriptome-wide association studies. Traditional models rely on genetic variants in close genomic proximity to the gene of interest to predict the genetic component of gene expression. Here we propose a novel approach incorporating distal genetic variants acting through gene regulatory networks, in line with the omnigenic model of complex traits. Results Using causal and coexpression Bayesian networks reconstructed from genomic and transcriptomic data, inference of gene expression from genotypic data is achieved through a two-step process. Initially, the expression level of each gene is predicted using its local genetic variants. The residual differences between the observed and predicted expression levels are then modeled using the genotype information of parent and/or grandparent nodes in the network. The final predicted expression level is obtained by summing the predictions from both models, effectively incorporating both local and distal genetic influences. Using regularized regression techniques for parameter estimation, we found that gene regulatory network-based gene expression prediction outperformed the traditional approach on simulated data and real data from yeast and human. This study provides important insights into the challenge of gene expression prediction for transcriptome-wide association studies. Availability github.com/guutama/GRN-TI .

Gutama Ibrahim Mohammad

Tom Michoel

Predicting the genetic component of gene expression using gene regulatory networks

预测基因调控网络中基因表达的遗传成分

Abstract Although multiple neural networks have been proposed for detecting secondary structures from medium-resolution (5–10 Å) cryo-electron microscopy (cryo-EM) maps, the loss functions used in the existing deep learning networks are primarily based on cross-entropy loss, which is known to be sensitive to class imbalances. To monitor and tune the performance of various loss functions for the secondary structure detection problem, we investigated five loss functions: cross-entropy, Focal loss, Dice loss, and two combined loss functions. Using a U-Net architecture in our DeepSSETracer method and a dataset composed of 1,355 box-cropped atomic-structure/density-map pairs, we found that a newly designed loss function that combines Focal loss and Dice loss provides the best overall detection accuracy for secondary structures. For β-sheet voxels, which are generally much harder to detect than helix voxels, the combined loss function achieved a significant improvement (an 8.8% increase in the F1 score) compared to the cross-entropy loss function and a noticeable improvement from the Dice loss function. This study demonstrates the potential for designing more effective loss functions for hard cases in the segmentation of secondary structures. The newly trained model was incorporated into DeepSSETracer 1.1 for the segmentation of protein secondary structures in medium-resolution cryo-EM map components. DeepSSETracer 1.1 can be integrated into ChimeraX, a popular molecular visualization software.

Yongcheng Mu

Thu Nguyen

Bryan Hawickhorst

Jiangwen Sun

The Combined Focal Loss and Dice Loss Function Improves the Segmentation of Beta-Sheets in Medium-Resolution Cryo-Electron-Microscopy Density Maps

结合焦点损失和Dice损失函数改善中分辨率冷冻电镜密度图中β片层分割效果

Abstract Motivation The Galaxy workflow system is an open-source platform supporting data-intensive research in life sciences, featuring a user-friendly web interface for complex analyses without extensive programming. It also offers a representational state transfer based API, enabling remote execution of specific tools. Galaxy supports similarity searches for nucleotide and amino acid sequences, with integrated tools like NCBI BLAST+ and DIAMOND. However, no specialized software currently exists for convenient use of NCBI BLAST+ and DIAMOND via the Galaxy API. Results blast2galaxy is a Python package that uses the Galaxy API to run sequence alignments with NCBI BLAST+ and DIAMOND as Galaxy-wrapped tools on compatible servers. It includes a command-line interface that mirrors the CLI of BLAST+ and DIAMOND and a high-level Python API for direct alignments from Python applications. The package relies on bioblend for communication with the Galaxy API. Availability and implementation blast2galaxy is available as open-source software under the MIT license. The source code is available on Github: https://github.com/IPK-BIT/blast2galaxy. It can be installed from the Python Package Index using “pip install blast2galaxy” or from the Bioconda channel using “conda install -c bioconda blast2galaxy”. Docker and Apptainer images are available and referenced in the documentation which is available under https://blast2galaxy.readthedocs.io

Patrick König

Anne Fiebig

Thomas Münch

Björn Grüning

Uwe Scholz

blast2galaxy: a CLI and Python API for BLAST+ and DIAMOND searches on Galaxy Servers

blast2galaxy：用于Galaxy服务器的BLAST+和DIAMOND搜索的CLI和Python API

Abstract Motivation Many evidences suggest that the subcellular localization of long-stranded non-coding RNAs (LncRNAs) provides key insights for the study of their biological function. Results This study proposes a novel deep learning framework, LncLSTA, designed for predicting the subcellular localization of LncRNAs. It firstly exploits LncRNA sequence, electron-ion interaction potential (EIIP), and nucleotide chemical property (NCP) as feature inputs. Departing from conventional k-mer approaches, this model employs a set of one-dimensional convolutional and maxpooling operations for dynamical feature aggregation. Furthermore, LncLSTA integrates a long-short term attention module with a Bi-LSTM network to comprehensively extract sequence information. In addition, it incorporates a TextCNN module to enhance accuracy and robustness in subcellular localization tasks. Experimental results demonstrate the efficacy of LncLSTA, showcasing its superior performance compared to other state-of-the-art methods. Notably, LncLSTA exhibits the transfer learning capability, extending its utility to predict the subcellular localization prediction of mRNAs, while maintaining consistently satisfactory prediction results. This research contributes valuable insights into understanding the biological functions of LncRNAs through subcellular localization, emphasizing the potential of deep learning approaches in advancing RNA-related studies. Availability The source code is publicly available at https://bis.zju.edu.cn/LncLSTA. Supplementary information Supplementary data are available at Bioinformatics online.

Kai Wang

Yueming Hu

Sida Li

Ming Chen

Zhong Li

LncLSTA: A versatile predictor unveiling subcellular localization of lncRNAs through long-short term attention

LncLSTA：一种通过长短期注意力揭示lncRNA亚细胞定位的多功能预测器

Abstract Motivation Molecular interaction networks are powerful tools for studying cellular functions. Integrating diverse types of networks enhances performance in downstream tasks such as gene module detection and protein function prediction. The challenge lies in extracting meaningful protein feature representations due to varying levels of sparsity and noise across these heterogeneous networks. Results We propose ICoN, a novel unsupervised graph neural network model that takes multiple protein-protein association networks as inputs and generates a feature representation for each protein that integrates the topological information from all the networks. A key contribution of ICoN is exploiting a mechanism called “co-attention” that enables cross-network communication during training. The model also incorporates a denoising training technique, introducing perturbations to each input network and training the model to reconstruct the original network from its corrupted version. Our experimental results demonstrate that ICoN surpasses individual networks across three downstream tasks: gene module detection, gene coannotation prediction, and protein function prediction. Compared to existing unsupervised network integration models, ICoN exhibits superior performance across the majority of downstream tasks and shows enhanced robustness against noise. This work introduces a promising approach for effectively integrating diverse protein-protein association networks, aiming to achieve a biologically meaningful representation of proteins. Availability The ICoN software is available under the GNU Public License v3 at https://github.com/Murali-group/ICoN.

Nure Tasnina

T M Murali

ICoN: Integration using Co-attention across biological networks

ICoN：利用生物网络间的协同注意力进行整合

短名	Bioinform. Adv.
Journal Impact	2.32
国际分区	MATHEMATICAL & COMPUTATIONAL BIOLOGY(Q2)

涉及主题	bioinformaticscomputational biology
出版信息	出版商: Oxford University Press，出版周期: ，期刊类型: journal，开源期刊: 是
基本数据	创刊年份: 2021，原创研究文献占比: ，自引率:， Gold OA占比:
期刊费用	$2620详情
平均审稿周期	8 weeks

Bioinformatics Advances

期刊引文格式

有 1 位以上作者的期刊

有 2 位作者的期刊

有 3 位作者的期刊

有 5 位以上作者的期刊

书籍引用格式

学位论文引用格式

网页引用格式

专利引用格式

手工熬夜修改参考文献？研飞自动匹配期刊，一键轻松成稿，支持 Word/WPS

投稿经验分享