mafft

high-performance multiple sequence alignment program

TLDR

Align sequences (auto strategy)

$ mafft --auto [input.fasta] > [aligned.fasta]

Fast alignment

$ mafft --retree 1 [input.fasta] > [aligned.fasta]

Accurate alignment

$ mafft --maxiterate 1000 --localpair [input.fasta] > [aligned.fasta]

Use multiple threads

$ mafft --thread [8] --auto [input.fasta] > [aligned.fasta]

Add sequences to existing alignment

$ mafft --add [new.fasta] [existing.fasta] > [combined.fasta]

SYNOPSIS

mafft [options] input > output

MAFFT (Multiple Alignment using Fast Fourier Transform) is a high-performance multiple sequence alignment program. It offers various algorithms balancing speed and accuracy.MAFFT is widely used in bioinformatics for aligning DNA, RNA, and protein sequences.

PARAMETERS

--auto

Auto-select strategy.

--maxiterate n

Number of iterations.

--localpair

Use L-INS-i (accurate).

--globalpair

Use G-INS-i.

--thread n

Number of threads.

--add file

Add to alignment.

--retree n

Number of tree-building iterations in progressive alignment. 1 is fast, 2 is default.

--reorder

Reorder output sequences by similarity.

--adjustdirection

Automatically detect and reverse-complement input sequences if needed.

--quiet

Suppress progress messages and warnings.

STRATEGIES

$ FFT-NS-1  - Fast, suitable for large datasets (--retree 1)
FFT-NS-2  - Default progressive method (--retree 2)
FFT-NS-i  - Iterative refinement (--maxiterate >0)
L-INS-i   - Most accurate, local alignment, <200 sequences (--localpair)
G-INS-i   - Global alignment (--globalpair)
E-INS-i   - For sequences with long unalignable regions (--genafpair)

CAVEATS

Memory scales with sequence count. Very long sequences may need adjustments. Auto mode selects strategy based on input size.

HISTORY

MAFFT was developed by Kazutaka Katoh and colleagues, first published in 2002. It's one of the most cited alignment tools in bioinformatics.