RAxML

RAxML
Name	RAxML
Developer	Alexandros Stamatakis et al.
Initial release	2006
Latest release	8.x / 9.x series
Programming language	C, C++
Operating system	Linux, macOS, Windows (via Cygwin/WSL)
License	GNU GPL / academic
Website	(see project pages)

RAxML is a software tool for maximum likelihood-based phylogenetic inference widely used in computational biology, systematics, and evolutionary genomics. It is designed for estimating large phylogenies from DNA, RNA, and protein sequence alignments and supports complex substitution models, bootstrap analysis, and parallel computation. RAxML is commonly cited in studies from molecular evolution, comparative genomics, and biodiversity assessment and is used alongside tools and resources from major institutions in the life sciences.

Overview

RAxML occupies a central role in phylogenetic analysis workflows employed by researchers at institutions such as European Bioinformatics Institute, Broad Institute, Max Planck Society, Smithsonian Institution, and Cold Spring Harbor Laboratory. It interfaces with alignment packages and databases like MAFFT, MUSCLE, Clustal Omega, GenBank, and RefSeq and complements downstream visualization and annotation platforms including FigTree, iTOL, Mesquite, and Dendroscope. RAxML is often compared with and used alongside other phylogenetic inference programs such as MrBayes, BEAST, PhyML, IQ-TREE, and PAUP*.

Features and Algorithms

RAxML implements maximum likelihood optimization under a variety of substitution models traditionally named for researchers and strains, such as Jukes–Cantor, Kimura 2-parameter, and Whelan and Goldman models, as well as mixed and partitioned models tailored for concatenated datasets. It integrates fast heuristics for tree search including rapid hill-climbing, subtree pruning and regrafting (SPR), and lazy subtree rearrangement strategies developed in the field of computational phylogenetics referenced by groups at University of California, Berkeley, Stanford University, and University of Oxford. For rate heterogeneity it supports discrete Gamma models popularized in work by researchers at University of Cambridge and University of Washington. RAxML also offers nonparametric and rapid bootstrap methods for assessing clade support that are commonly benchmarked against approaches used by teams at University of Tokyo and University of Edinburgh.

Implementation and Performance

RAxML is implemented in C/C++ with performance optimizations informed by parallel computing research from laboratories at Lawrence Berkeley National Laboratory, Argonne National Laboratory, and Sandia National Laboratories. It provides shared-memory parallelism via pthreads and OpenMP, and distributed-memory parallelism via MPI implementations used on clusters at CERN and national supercomputing centers like Oak Ridge National Laboratory. Performance tuning leverages CPU vector instructions similar to strategies from chip vendors such as Intel and AMD, and scales on architectures produced by IBM and ARM Holdings. Benchmarks published by collaborative teams at University of Tübingen and University of Vienna show RAxML performing competitively on large alignments typical of projects at European Molecular Biology Laboratory.

Usage and Input/Output Formats

RAxML accepts common alignment formats used in projects at National Center for Biotechnology Information, including PHYLIP and FASTA formats produced by tools like EMBOSS and Seqotron, and supports partition files and model specifications compatible with workflows at Wellcome Sanger Institute. Output includes maximum likelihood trees in Newick format used by visualization projects such as Archaeopteryx and tabular log files that integrate with pipeline managers developed at Broad Institute and European Bioinformatics Institute. RAxML can be invoked from command-line environments common at University of California, San Diego and integrates into workflow systems like Snakemake and Nextflow used by computational groups at ETH Zurich.

Applications and Case Studies

RAxML has been employed in large-scale phylogenomic studies led by consortia such as the Tree of Life Web Project and biodiversity initiatives supported by Botanical Gardens Conservation International. Case studies include vertebrate phylogeny reconstructions coordinated by researchers at Harvard University, microbial community phylogenetics in projects run by Joint Genome Institute, and viral evolution studies in collaborations with World Health Organization and Centers for Disease Control and Prevention. RAxML analyses feature in conservation genomics reports associated with International Union for Conservation of Nature assessments and in paleogenomics projects involving teams from Natural History Museum, London.

History and Development

RAxML development traces to work by Alexandros Stamatakis and collaborators with ties to research groups at Heidelberg University and subsequently incorporated contributions from developers associated with Technical University of Munich, University of Freiburg, and international collaborators at University of Athens. Over successive releases the codebase incorporated algorithmic advances presented at conferences such as the International Conference on Research in Computational Molecular Biology and Society for Molecular Biology and Evolution annual meetings. Development and maintenance have been influenced by funding and infrastructure from institutions including the German Research Foundation and European Commission frameworks.

Licensing and Availability

RAxML is distributed under open-source terms historically aligned with the GNU General Public License and academic use clauses embraced by many university groups including University of Heidelberg and Max Planck Institute for Informatics. Precompiled binaries and source code have been hosted on project pages and mirrors maintained by organizations such as European Bioinformatics Institute and community repositories supported by GitHub and Bitbucket. Commercial entities often coordinate with academic licensors and technology transfer offices at institutions like Stanford University when integrating RAxML into proprietary pipelines.

Category:PhylogeneticsCategory:Bioinformatics software