Publications by authors named "Alexey M Kozlov"
Article Synopsis
- * Research shows that by adjusting specific numerical thresholds in tools like RAxML-NG and IQ-TREE, tree inference times can be significantly improved without compromising the accuracy of the resulting trees.
- * The study provides access to datasets and scripts used in the research, allowing others to replicate or build upon their findings via the provided links.
Article Synopsis
- Phylogenetic networks are used to model complex evolutionary scenarios, but existing methods struggle with high computational demands and small datasets, especially when considering incomplete lineage sorting (ILS).
- The team introduces NetRAX, a maximum likelihood tool that simplifies network inference by avoiding ILS complications, utilizing efficient tree likelihood computations, and outputting results in Extended Newick format.
- NetRAX performs well on simulated data, providing accurate network inference quickly, and is available for use under an open-source license on GitHub.
Article Synopsis
- Phylogenetic inference is increasingly performed on powerful computing systems, leading to challenges in ensuring parallel fault tolerance for tools like RAxML-NG.
- The study explores necessary software modifications to achieve fault tolerance, revealing that the added recovery mechanisms lead to a minimal slow down of about 1.7 times on large datasets, but ensure reliable recovery from multiple types of failures.
- The updated RAxML-NG with fault tolerance is publicly available for use, promoting easier access to improved phylegenetic analysis methods.
Article Synopsis
- Many daily publications focus on analyzing SARS-CoV-2 data, including phylogenetic studies available on nextstrain.org.
- The authors discuss challenges in creating reliable phylogenies due to a high number of virus sequences but a low number of mutations, making it tough to draw clear evolutionary connections.
- They conclude that while phylogenetic methods can offer some insights into COVID-19's evolution and spread, researchers should interpret results with caution, especially when using standard analysis tools.
Article Synopsis
- Inferring phylogenetic trees for individual gene families is challenging due to short alignments and ineffective substitution models, so methods that incorporate species tree information are necessary.
- GeneRax is introduced as the first maximum likelihood software that considers both sequence-level and gene-level events, such as duplications and transfers, for phylogenetic inference.
- In simulations, GeneRax accurately infers trees in 90% of cases and is the fastest method for empirical data, effectively performing large-scale analyses, as demonstrated by its ability to process 1,099 Cyanobacteria families in just 8 minutes.
Article Synopsis
- Advances in microfluidics and low sequencing costs have revolutionized single-cell sequencing technology, allowing for the analysis of thousands to millions of cells in one experiment.
- This rapid data generation presents unique challenges in data science, which the text identifies as central to the future of single-cell biology.
- The article provides an overview of eleven key challenges, including motivating research questions and open problems, making it relevant for both experienced researchers and newcomers to the field.
Article Synopsis
- Researchers created a faster and more memory-efficient version of the transfer bootstrap expectation (TBE) method for phylogenetic analysis, addressing limitations of the original, resource-heavy tool.
- Their new implementation can be up to 480 times quicker and uses significantly less memory, making it better for large datasets.
- This optimized TBE method has been integrated into existing tools and is available for public use under an open-source license.
Article Synopsis
- ModelTest-NG is a new tool designed from the ground up to improve upon jModelTest and ProtTest, which are used for selecting model substitutions for nucleotide and amino acid sequences.
- It is significantly faster—one to two orders of magnitude—than its predecessors while maintaining equal accuracy and adds new features like bias correction and automatic processing.
- ModelTest-NG is open source and can be accessed under the GNU GPL3 license at its GitHub page.
Article Synopsis
- Researchers developed a new model for amino acid sequence evolution that incorporates protein structure, which is often overlooked despite its importance.
- This "structurally aware" model uses an expanded alphabet to describe amino acids along with their side-chain configurations, taking into account geometric patterns and dihedral angles.
- The new model outperforms traditional models in estimating evolutionary divergence and reconstructing ancestral states, highlighting the significance of side-chain geometry for understanding protein folding and function in evolutionary biology.
Article Synopsis
- Phylogenies are crucial for biological research and have applications in fields like biotechnology and medicine, but finding optimal trees using maximum likelihood methods is computationally challenging.
- RAxML-NG is a new implementation that improves upon the previous RAxML/ExaML algorithm, offering better accuracy, speed, and new features while analyzing complex datasets.
- The RAxML-NG code is accessible under GNU GPL, with a web service available for users, and additional supplementary data is provided online.
Phylogenomics and the evolution of hemipteroid insects.
Kevin P Johnson Christopher H Dietrich Frank Friedrich Rolf G Beutel Benjamin Wipfler Alexey M Kozlov
Proc Natl Acad Sci U S A
December 2018
Article Synopsis
- Hemipteroid insects, which make up over 10% of insect diversity, are important in ecosystems but their evolutionary relationships have been unclear in past studies.
- Recent phylogenomic analyses of 193 hemipteroid insect samples offer a clearer phylogeny, confirming the monophyly of the three main orders: Psocodea, Thysanoptera, and Hemiptera, and suggesting Thysanoptera is closely related to Hemiptera.
- The study also indicates that hemipteroid insects began diversifying over 365 million years ago and discusses the impact of these findings on understanding insect evolution and traits.
Article Synopsis
- Coalescent- and reconciliation-based methods are increasingly used to infer species phylogenies from genomic data, but current tools are inefficient for processing large multiple sequence alignments (MSAs).
- The new tool, ParGenes, allows researchers to simultaneously conduct model testing and maximum likelihood inference on thousands of MSAs, significantly speeding up the analysis.
- ParGenes has been tested successfully, processing over 20,000 phylogenetic gene trees in just 28 hours on a supercomputing cluster, and is available for public use on GitHub.
Article Synopsis
- Next generation sequencing (NGS) technologies have resulted in an overwhelming amount of molecular sequence data, particularly complicating the field of metagenetics which deals with identifying sequences from various microbial environments.
- Traditional phylogenetic placement methods, like EPA and PPLACER, struggle with scalability due to advancements in NGS, prompting the development of a faster, more efficient tool called EPA-NG.
- EPA-NG can operate on both shared and distributed memory systems, showcasing impressive performance by processing 1 billion metagenetic reads within 7 hours on a computing cluster, while significantly outpacing previous algorithms.
Article Synopsis
- Creating a comprehensive and sustainable plant tree of life is becoming possible but faces challenges due to issues with current data integration and accessibility for non-experts.* -
- Existing phylogenetic trees are often static and quickly outdated, highlighting the need for a collaborative and adaptable framework for integrating DNA data and conducting phylogenetic analyses.* -
- The scientific community should focus on developing user-friendly interfaces for data access, regular updates of phylogenetic trees, and enhancing data quality through user feedback to achieve effective global phylogenetic synthesis.*
Article Synopsis
- - Public databases often contain molecular sequences annotated by the original authors, leading to potential mislabeling and errors that are difficult to detect, which can negatively impact metagenetic studies.
- - The research introduces SATIVA, a method that uses phylogenetic analysis to automatically identify and correct taxonomically mislabeled sequences, achieving high accuracy rates in both identification and correction.
- - Analysis of popular microbial reference databases reveals a significant presence of mislabels, ranging from 0.2% to 2.5%, and SATIVA provides a tool for exploring better taxonomic classifications, specifically for Cyanobacteria.
Article Synopsis
- Phylogenetic analysis is becoming crucial in medical and biological research, fueled by the rapid growth of datasets from next-generation sequencing.
- ExaML version 3 has been developed as a high-performance tool for inferring phylogenies from large datasets, including whole-transcriptome and whole-genome alignments, utilizing supercomputers.
- The latest version introduces enhancements like new substitution models, a novel load balance algorithm for better performance, and is optimized for Intel MIC-based hardware platforms.