Bioinformatics

GenomeDelta: detecting recent transposable element invasions without repeat library

1. Riccardo Pianezza
2. Anna Haider
3. Robert Kofler

Reviewed by Arcadia Science

Limitations of current machine learning models in predicting enzymatic functions for uncharacterized proteins

1. Valérie de Crécy-Lagard
2. Raquel Dias
3. Nick Sexson
4. Iddo Friedberg
5. Yifeng Yuan
6. Manal A Swairjo

Reviewed by Arcadia Science

Exploring the accuracy of ab initio prediction methods for viral pseudoknotted RNA structures

1. Vasco Medeiros
2. Jennifer M. Pearl
3. Mia Carboni
4. Ece Er
5. Stamatia Zafeiri

Reviewed by PREreview

Design of linear and cyclic peptide binders of different lengths from protein sequence information

1. Qiuzhen Li
2. Efstathios Nikolaos Vlachos
3. Patrick Bryant

Reviewed by PREreview

Deep Geometric Framework to Predict Antibody-Antigen Binding Affinity

1. Nuwan Bandara
2. Dasun Premathilaka
3. Sachini Chandanayake
4. Sahan Hettiarachchi
5. Vithurshan Varenthirarajah
6. Aravinda Munasinghe
7. Kaushalya Madhawa
8. Subodha Charles

Reviewed by PREreview

FusOn-pLM: A Fusion Oncoprotein-Specific Language Model via Focused Probabilistic Masking

1. Sophia Vincoff
2. Shrey Goel
3. Kseniia Kholina
4. Rishab Pulugurta
5. Pranay Vure
6. Pranam Chatterjee

Reviewed by PREreview

PPIscreenML: Structure-based screening for protein-protein interactions using AlphaFold

1. Victoria Mischley
2. Johannes Maier
3. Jesse Chen
4. John Karanicolas

• Curated by eLife

  eLife assessment

  This study explores simple machine learning frameworks to distinguish between interacting and non-interacting protein pairs, offering solid computational results despite some concerns about dataset generation. The authors demonstrate a modest improvement in AlphaFold-multimers' ability to differentiate these pairs. Using a simple yet sound approach, this work is a valuable contribution to the challenging problem of reconstructing protein-protein interaction networks.

Reviewed by eLife

PhysiCell Studio: a graphical tool to make agent-based modeling more accessible

1. Randy Heiland
2. Daniel Bergman
3. Blair Lyons
4. Grant Waldow
5. Julie Cass
6. Heber Lima da Rocha
7. Marco Ruscone
8. Vincent Noël
9. Paul Macklin

• Curated by GigaByte

  Editors Assessment:

  This paper presents a new tool to make using PhysiCell easier, which is an open-source, physics-based multicellular simulation framework with a very wide user base. PhysiCell Studio is a graphical tool that makes it easier to build, run, and visualize PhysiCell models. Over time, it has evolved from being a GUI to include many additional functionalities, and can be used as desktop and cloud versions. This paper outlines the many features and functions, the design and development process behind it, and deployment instructions. Peer review improved the organisation of the various repositories and adding both a requirements.txt and environment.yml files. Looking to the future the developers are planning to add new features based on community feedback and contributions, and this paper presents the many code repositories if readers wish to contribute to the development process.

  This evaluation refers to version 1 of the preprint

Reviewed by GigaByte

Low-coverage whole genome sequencing for a highly selective cohort of severe COVID-19 patients

1. Renato Santos
2. Víctor Moreno-Torres
3. Ilduara Pintos
4. Octavio Corral
5. Carmen de Mendoza
6. Vicente Soriano
7. Manuel Corpas

• Curated by GigaByte

  Editors Assessment:

  Many studies have explored the genetic determinants of COVID-19 severity, these GWAS studies using microarrays or expensive whole-genome sequencing (WGS). Low-coverage WGS data can be imputed using reference panels to enhance resolution and statistical power while maintaining much lower costs, but imputation accuracy is difficult to balance. This work demonstrates how to address these challenges utilising the GLIMPSE1 algorithm, a less resource-intensive tool that produces more accurate imputed data than its predecessors. Generating a dataset containing 79 imputed low-coverage WGS samples from patients with severe COVID-19 symptoms during the initial wave of the SARS-CoV-2 pandemic in Spain. The validation of this imputation and filtering process shows that GLIMPSE1 can be confidently used to impute variants with minor allele frequency up to approximately 2%. After peer review the authors clarified and provided more validation and statistics and figures to help convince this approach was valid. This work showcasing the viability of using low-coverage WGS imputation to generate data for the study of disease-related genetic markers, alongside a validation methodology to ensure the accuracy of the data produced. Helping inspire confidence and encouraging others to deploy similar approaches to other infectious diseases, genetic disorders, or population-based genetic studies. Particularly in large-scale genomic projects and resource-limited settings where sequencing at higher coverage could prove to be prohibitively expensive.

  This evaluation refers to version 1 of the preprint

Reviewed by GigaByte

LRTK: a platform agnostic toolkit for linked-read analysis of both human genome and metagenome

1. Chao Yang
2. Zhenmiao Zhang
3. Yufen Huang
4. Xuefeng Xie
5. Herui Liao
6. Jin Xiao
7. Werner Pieter Veldsman
8. Kejing Yin
9. Xiaodong Fang
10. Lu Zhang

Reviewed by GigaScience