MAFFT

MAFFT
Name	MAFFT
Developer	Kazutaka Katoh
Released	2002
Latest release version	7.x
Operating system	Unix-like, MacOS, Windows
License	GPL

MAFFT MAFFT is a widely used multiple sequence alignment program for biological sequences. It provides progressive, iterative, and FFT-based alignment strategies suitable for datasets ranging from small protein families to large-scale genomic projects. Its development intersects with major bioinformatics resources and institutions involved in sequence analysis, high-performance computing, and comparative genomics.

Overview

MAFFT offers a suite of algorithms for constructing alignments of nucleotide and protein sequences and supports both local and global alignment paradigms. The package is relevant to groups using tools from European Bioinformatics Institute, National Center for Biotechnology Information, Genome Institute at Washington University, and projects like 1000 Genomes Project and ENCODE Project. Researchers in laboratories affiliated with Max Planck Society, Wellcome Trust Sanger Institute, Broad Institute, and Cold Spring Harbor Laboratory frequently compare MAFFT performance with alternatives such as Clustal Omega, MUSCLE, T-Coffee, ProbCons, and PRANK.

Features and Algorithms

MAFFT implements fast Fourier transform (FFT)-accelerated routines and iterative refinement strategies derived from progressive alignment frameworks popularized by software like Clustal W and Clustal X. It contains modes including FFT-NS-1, FFT-NS-2, L-INS-i, G-INS-i, and E-INS-i, each tuned for different evolutionary scenarios comparable to methods used in BLAST, HMMER, and PhyML. The program integrates pairwise alignment scoring, consistency-based scoring reminiscent of approaches in T-Coffee, and guide tree construction often informed by distance measures akin to those in Neighbor-joining and algorithms implemented in MEGA. MAFFT also offers options for profile-to-profile alignment similar to interfaces in ExPASy, and supports structural alignment inputs as employed in studies involving Protein Data Bank entries and pipelines used by Rosetta.

Usage and Command-Line Options

MAFFT is invoked primarily from a command line environment on systems such as Linux, macOS, and Microsoft Windows via terminal emulators or scripting frameworks used in laboratories at institutions like Stanford University, Massachusetts Institute of Technology, and University of California, Berkeley. Common options include model selection flags for scoring matrices comparable to those found in BLOSUM62 and PAM250 usage in other programs, gap open and gap extension parameters akin to settings in Needleman–Wunsch implementations, and iterative refinement switches that echo choices in RAxML and IQ-TREE workflows. Batch processing is frequently orchestrated alongside workflow managers such as Snakemake, Nextflow, and Cromwell in large consortia like Human Cell Atlas and Earth Microbiome Project.

Performance and Accuracy

Benchmarking studies compare MAFFT against tools evaluated in publications from journals like Nature Methods, Bioinformatics (journal), and Genome Research. Performance assessments reference datasets curated by projects such as BAliBASE, SABmark, and HomFam, and consider runtime scaling on hardware from vendors like Intel, AMD, and accelerators from NVIDIA. Accuracy comparisons involve downstream phylogenetic reconstruction with programs such as RAxML-NG, MrBayes, and BEAST, and consider impacts on analyses performed in consortia including Tree of Life Web Project. MAFFT's FFT-based schemes often yield favorable runtime for large alignments, while iterative accuracy-focused modes are competitive for alignments used in structural inference and comparative genomics.

Implementations and Integrations

MAFFT is distributed as standalone binaries and source code, integrated into bioinformatics pipelines hosted on platforms such as GitHub, Bioconda, and Debian. It is wrapped by web services and portals comparable to interfaces at EMBL-EBI and integrated into analysis suites like Ugene, Seaview, and Galaxy. MAFFT outputs are commonly consumed by downstream tools for phylogenomics, orthology inference, and variant interpretation in ecosystems involving OrthoMCL, eggNOG, InterProScan, and visualization in Jalview or AliView.

History and Development

Development began under the stewardship of Kazutaka Katoh with contributions from collaborators associated with institutions such as Kyoto University, Tohoku University, and research groups connected to RIKEN. Early releases emphasized FFT acceleration and practical speed-ups for protein alignment tasks, with successive versions adding iterative refinement, structural awareness, and large-scale strategies to address needs identified in community efforts like Human Genome Project follow-ups and comparative initiatives at National Institutes of Health. Maintenance, feature additions, and platform packaging have involved community contributors distributed through repositories and package managers used by research groups at University of Tokyo and international collaborators.

Applications and Limitations

MAFFT is applied across domains including molecular phylogenetics in studies published by researchers at Scripps Research, University of Cambridge, and Imperial College London; comparative transcriptomics in projects associated with The Broad Institute; and metagenomics workflows for initiatives like Tara Oceans. Limitations include sensitivity to highly divergent sequence sets and alignment uncertainty that can affect downstream inferences in pipelines employing BEAST 2 or MrBayes; users often complement MAFFT with masking tools and guide-tree strategies used in protocols from groups at European Molecular Biology Laboratory and Wellcome Sanger Institute. Advanced use cases may require combining MAFFT with structural alignment programs such as MUSTANG or machine-learning approaches emerging from labs at Google DeepMind and Microsoft Research.

Category:Bioinformatics software