Biology Reference
In-Depth Information
Chapter 8
MAFFT: Iterative Refinement and Additional Methods
Kazutaka Katoh and Daron M. Standley
Abstract
This chapter outlines several methods implemented in the MAFFT package. MAFFT is a popular multiple
sequence alignment (MSA) program with various options for the progressive method, the iterative refine-
ment method and other methods. We first outline basic usage of MAFFT and then describe recent practical
extensions, such as dot plot and adjustment of direction in DNA alignment. We also refer to MUSCLE,
another high-performance MSA program.
Key words
Multiple sequence alignment, Iterative refinement, Fast Fourier transform, Metagenome,
Protein structure
1
Introduction
MAFFT [
1
] is a general purpose multiple sequence alignment
(MSA) program for nucleotide or amino acid sequences. Due to
its high performance [
2
-
6
], MAFFT is becoming popular in recent
years. It has several different options depending on the size and
type of alignment problem. In this chapter, first we outline various
MAFFT options. Then, we describe several practical new features
for web service and command-line versions. As a possible direction
for future development, we discuss our ongoing efforts to use
structural information in MSA calculations. We also clarify the
limitation of MAFFT in actual analyses, by showing typical inap-
propriate usage. In addition, we refer to MUSCLE [
7
,
8
], another
high-performance MSA program.
2 Basic Algorithms
MAFFT has various options as discussed below. Generally, there is a
trade-off between speed and accuracy. However, there are excep-
tional situations (
see
Subheading
6
) that should be carefully consid-
ered when selecting an option to apply. Like most MSA methods,
