Vol.9 no.
6 1993
CABIOS
DCSE, an interactive tool for sequence
alignment and secondary structure research
Peter De Rijk and Rupert De Wachter1
Abstract
DCSE provides a user-friendly package for the creation and
editing of sequence alignments. The program runs on different
platforms, including microcomputers and workstations. Apart
from available hardware, the program is not limited in the size
of the alignment it can handle. It deviates more from classical
text editors than other available sequence editors because it uses
a different approach towards editing. It shifts characters or
entire blocks of aligned characters, rather than inserting or
deleting gaps in the sequences. Alignment of a new sequence
to an existing alignment is partly automated. Although DCSE
can be used on protein sequence alignments, it is especially
targeted at the examination ofRNA. The secondary structure
for every sequence can be incorporated easily in the alignment.
DCSE also has extensive built-in support for finding and
checking secondary structure elements. A sophisticated system
of markers allows notation of special positions in an alignment.
This system can be used to store information such as the position
of hidden breaks, introns and tertiary structure interactions.
Introduction
Comparison of biological macromolecules is gaining in
importance in studies on the structure and evolution of those
molecules. Due to the development of rapid methods for
sequence determination, there has been an explosive growth
in published sequence data. Since a sequence alignment forms
the basis of most comparative studies, much effort has been
put into programs that aid in the creation of multiple sequence
alignments.
Several programs were developed to automate totally the
creation of an optimal alignment (reviewed by Chan et al.,
1992). However, the creation, maintenance and exploration of
a large, meaningful alignment remains a challenge, and the
advantages of an interactive approach have been advocated by
several authors (Rechid etal, 1989; Schuler etal, 1991;
Depiereux and Feytmans, 1992). Indeed, multiple sequence
alignment programs use a theoretical mathematical model to
define what is considered an optimal alignment. The rules
followed by these programs are, by necessity, rather limited
Departement Biochemie, Universiteit Antwerpen {VIA), Universiteitsplein 1, environment to create and maintain sequence alignments. It is
B-2610 Antwerpen, Belgium
l
To whom reprint requests should be sent
Oxford University Press
Pages 735-740
and strict. This makes it difficult to incorporate extra
information such as a description of secondary structure into
Downloaded from http://bioinformatics.oxfordjournals.org/ at University of Michigan on September 19, 2015
the alignment process. As a result the alignments produced by
these programs are not always optimal from a biological point
of view. Also, the number of sequences that can be aligned
by these programs is usually quite limited.
Another approach to the problem of making an alignment
is manual editing. Ordinary text editors are not very well suited
for this task, mainly due to the limited number of characters
that can be placed on one line. Several multiple sequence
alignment editors have been developed in response to these
problems (Stockwell and Petersen, 1987; Faulkner and Jurka,
1988; Thirup and Larsen, 1990; Olsen et al. ,1991; Parry-Smith
and Attwood, 1991; Clark, 1992). Most of these editors seem
to be aimed primarily at protein sequences, and though they
can be used on nucleic acid sequences, they do not take into
account the peculiarities involved in aligning a large set of
structural nucleic acid sequences. They are modelled after
existing full-screen text editors but overcome the line length
problem, and feature functions typically used on sequence
alignments. These editors use the classical methods for editing
text to align sequences. Alignment is achieved by inserting or
deleting gap symbols. While these methods are very suitable
for entering and editing text, a different approach seems more
suited for editing an alignment. Biopolymer sequences have
been experimentally determined, and should not be changed
unless errors are discovered. Alignment does not involve
changing, but shifting characters and groups of characters in
the sequences.
In the case of molecules such as ribosomal RNA, alignment
of more variable areas can be guided by secondary structure
information. A sequence editor for these molecules should
provide an easy and straightforward method to incorporate this
information in the alignment. It should also provide tools to
locate structural elements, and to check whether the encoded
structure is consistent.
DCSE (Dedicated Comparative Sequence Editor) originated
from the need to maintain a rapidly growing alignment of small
subunit ribosomal RNAs (De Rijk etal., 1992). This file
currently contains 2000 sequences and has 4760 alignment
positions. DCSE was developed from the start to cope with large
alignments. It provides a user-friendly, menu-driven
especially suited for the study of sequences following a common
secondary structure pattern.
735
P.De Rjjk and R.De Wachter
System and methods
DCSE was written following the ANSI specification of the C
programming language. However, some parts are necessarily
platform specific. These parts lie mainly in routines interfacing
with the operating system, such as screen manipulation,
keyboard and filing system routines. DCSE has been compiled
for the following environments: VMS for VAXstations, Ultrix
on DECstations, DOS on IBM-compatible PCs and RISC OS
on the Acorn Archimedes range.
Under VMS, DCSE has been compiled using VAX C. It uses
ANSI escape sequences to control the display and the SMG
routines for keyboard input. The version on DECstations was
compiled with Ultrix C, and uses Curses for both screen output
and keyboard input. Implementations on Turbo C for DOS and
Desktop C for RISC OS use commands provided by these
packages to handle the necessary screen output and keyboard
input.
Algorithms
Editing
DCSE looks at an alignment as if it were an abacus. The
alignment has a rod or sequence line for every organism. All
rods have the same length, which is set by the number of
positions. This number can be reduced by removing positions
that are empty in all sequence lines, or increased by inserting
new positions. Every rod has a fixed number of beads
(characters) in a fixed order. The number of characters is
smaller than the number of positions, so every position contains
either a nucleotide symbol or a gap symbol. This way the
characters can be shifted. If a character is pushed leftward or
rightward, and makes contact with another one, the latter is
pushed in the same direction. In this way the fixed order of
characters, or primary structure, will always remain correct.
Just as in an ordinary screen editor, DCSE uses the screen
as a window on a part of the alignment. This window can be
moved in several ways to display other parts of the alignment.
The screen can be split to show two windows on different parts
of the alignment. It also features a pointer to show the current
position in the alignment. This pointer can be moved by the
arrow keys, and the window will scroll appropriately in order
to keep the pointer on the screen. It is used as a finger, which
can push characters leftward or rightward. The pointer cannot
only push characters, it can also move characters to the other
end of a gap, get a character from the other end of a gap, or
move a continuous block of characters to either side. The pointer
can also be resized so that it covers a number of sequences and
positions. The resized pointer can perform the same actions as
the small one, but all characters covered by the pointer will
keep their relative positions during the process.
The order of the sequence lines is not rigid. One or more
lines can be locked in a given position on the screen, while
736
the rest of the alignment can still be moved up- or downward.
This makes it easy to compare one sequence to several others,
or to rearrange the order of the sequences temporarily.
Checking the primary structure
DCSE uses reference files to store a version of the sequences
that is never changed. The reference file can also contain extra
information about the sequence, such as the taxonomic position
of the organism, and a literature reference. The sequences in
the reference file are generally not edited, and thus remain
correct. DCSE can check whether the sequence in an alignment,
Downloaded from http://bioinformatics.oxfordjournals.org/ at University of Michigan on September 19, 2015
disregarding the gaps, is equal to the corresponding sequence
in a reference file. It does this by aligning the two sequences.
DCSE will show the differences on the screen, and offer to
correct them. Users can decide to make a correction themselves,
if the alignment might be disturbed by automatic insertion or
deletion of characters.
Aligning sequences
One sequence can be automatically aligned to another one by
DCSE. The alignment algorithm will create an array the size
of one sequence. Corresponding positions in the other sequence
will be stored in this array. For the creation of the array it uses
a combination of two methods. The program starts by
comparing the two sequences using a recursive method.
Subsequences of specified size that appear in both sequences
are matched. The unmatched stretches between two matched
subsequences are analysed using the same method with a smaller
block size. When the subsequence size reaches a certain
minimum, corresponding positions in the remaining stretches
are matched using an algorithm that works by minimizing an
'alignment distance' between the stretches (Sellers, 1974). This
distance is calculated by adding a penalty for every mismatch
or gap. Gaps are penalized using affine gap costs (Spouge,
1991).
The sequence that has served as a reference will not be
changed by the alignment routine. The characters of the other
sequence will be shifted relative to this sequence in order to
reflect the calculated correspondence array. However, if the
newly aligned sequence contains an insertion in a spot where
the reference sequence does not have a corresponding gap, this
can not be properly accommodated in the alignment. DCSE's
alignment routine can handle this situation in three ways. It can
create a global insert in the entire alignment, or it can carry
out the insert by pushing the surrounding characters in the newly
aligned sequence aside, thereby possibly disrupting the
alignment locally. The other option is to leave the insertion out.
This option will produce an error in the primary structure, which
can be detected easily later on by the primary structure checking
routine. This will leave it up to the user to decide whether a
global insert should be created, or whether the problem can
be solved by a local sequence realignment.
 Dedicated Comparative Sequence Editor

Markers the identification of secondary structure elements. For

In its simplest form, a marker is a position in the alignment
that can be given a name. The user can go to a marked position
by selecting its name, or by going to the next or previous
marker. A whole range of options and functions to control
markers provides a flexible system to note special positions in
an alignment. A marker can be located at a certain alignment
or sequence position. A sequence name can be associated with
a marker. In this case, subsequent selection of this marker will
move the pointer to the given position in this particular
sequence. Otherwise it will move to the position on the current
homologous sequences with sufficiently similar secondary
structures, one helix numbering line will suffice. Three such
lines are present in our SSU RNA database, describing the
secondary structure of eukaryotic, prokaryotic and mito-
chondrial sequences. The coding system allows practically all
possible structures, including pseudoknots.
The secondary structure indicated can also be tested. DCSE
uses the first helix numbering line to make a table containing
the approximate positions of both strands of each helix. When
checking the secondary structure, it searches for the first

Downloaded from http://bioinformatics.oxfordjournals.org/ at University of Michigan on September 19, 2015

sequence line. A marker can also refer to two correlated
positions. A range of markers can be inactivated or activated.
A range is selected by typing in a part of the name. All markers
containing this part in their name will be subsequently activated
or inactivated. This feature can be used to create different groups
of markers that can be used and saved separately.
During editing, markers can be used to keep track of
interesting subsequences. Marker files can also be used to keep
all sorts of extra positional information at hand. It could be used,
for example, to mark the position of hidden breaks, introns or
tertiary structure elements. Markers can be saved in, or loaded
from, a marker file. A marker file is an ASCII text in a simple
format, and it can easily be generated by other programs or
functions.
Indication and checking of RNA secondary structure
In each sequence line the symbols for nucleotides or gaps
alternate with positions that are either blank or contain a special
symbol. These can be special characters which aid in the
alignment process. However, their main use is to delimit
secondary structure elements. These symbols move together
with the characters. Their use is illustrated in Figure 1. 'Helix
numbering' lines are intercalated between the sequences to allow
occurrence of a helix strand starting from the pointer. The helix
position table is checked to find out which helix this strand is
a part of. The opposite strand is found, and is checked for
complementarity. When the program encounters an error in the
proposed secondary structure it stops, gives an appropriate error
message, and indicates the location of this error in the alignment.
There are options to check only one helix, and to check whether
helices can be extended. DCSE has a function to copy a
secondary structure from another sequence line. In order to find
the structure of a newly aligned sequence, the secondary
structure description of a closely related organism can be copied,
and its consistency checked.
Implementation
There are few limitations to the size of the alignments DCSE
can handle. As memory is allocated dynamically, the only limit
to the size and number of sequences that can be loaded is
available memory. DCSE can even be used to edit an alignment
larger than memory relatively easily; it allows the user to load
and work on just a part of the total alignment. Some of the
changes made to this part, such as global inserts, are applied
to the entire alignment, and not only the part loaded.
Editors can be fairly complicated programs to use. DCSE

(G A U{G A ) G U U*G C G G A ) G U G A [ U C C G C*C U G G)U U A A U|C C { A ) A G ) U U A[A A C { A ) A U C) Organism 1

[G A U{G A } G A U'G C ( G ) G A)G U G A [ U C ( A ) G V'C U G G)U U A A U|C C - A G]U U A[A U C { A ) A U C) Organism 2
Helix numbering

Organism 1 u Organism 2 ; * AU
u G C
G G
G U c
u C A
c G
r- * U C A GU G
UCC8C U U
U 2
u
u A GGCG u
A G G CG A A A
G
A G U
A A - A
U
C
A
A
A
A
U A - U
G C
c

Fig. 1. Illustration of secondary structure symbols. Two imaginary RNA sequences and drawings of corresponding secondary structures are shown. In the linear
representation square brackets designate the beginning and end of one strand of a helix and a circumflex separates adjacent helix strands. Braces are used to
indicate the beginning and end of an internal loop or bulge loop, and a base taking part in a non-standard pair is enclosed in parentheses. The 'Helix numbering'
line identifies the helices.

737
P.De R p and R.De Wachter

ros:0<!ba7D Splrocnaeta litoralls arteub.ali

tries to provide a user-friendly interface. Menus tend to be
preferred by users who are new to a program, since they show
the current range of options available. However, the more
experienced user can work faster and easier when often-used
functions are available at the stroke of a key. DCSE therefore
offers both possibilities. All functions, except the basic moving
and shifting routines, can be called from a hierarchical menu
system. Menu items can be selected using arrow keys. Most
functions have keyboard shortcuts as well. A help screen can
be called to list the meaning of most important keys and a
constantly updated help line shows the purpose of the currently
selected menu item or of certain keys. A user manual has also
Change colours: 16:G
i Enter <colourcode(s)>:<characters you uant to have this colour> L
c : -CAG-IIGAGGG-ACGAAAGd 1AGGG-1 iCl ICGAACCGGAIIIIAG-AI IACCC-GG-GI IAGUCCI IAGOiGt I-AA
s i : -C6C-IJGAGGC-GNNA--AAGCG! 'GGG-GAGCAAACAGGA! ihAG-MMCCC-i W-GUAGI iCCftCGCNGU-AA
: CAC I1GAGGU GCGA AAGCG GGG GAGl'AAAl ftGGAIIliAG Ai:ACCC G GIIAGIXCA CGCCGll AA
: CGC UCAIIGC ACGA AAGCG IGGG GAGCAflACAGGAI IIIAG AliACCC UG GHAGIICCA CGCCCU AA
s :-CGC-IICAUGll-GCNA--AAGC&iGGG-GAGCAAACAGGAUUAG-AHACCC HG-GIIAGIICCA--CGCIIGII-AA
p :-CAC-UCAGAU-GCGA--AAGCG:lGGG GAGCAAACAGGAUUAG-AUACCC UC-GUCGUCCA-CGCCGU-AA
u : CGC IIKAGGII ACNA AGGCG 1GGG IIAGCGAACGGGAIIIAG AlAtXC CG GHAGIICCA CGCAGH AA
s i : CGC llhAGGC GMNAMSC6UC6G IIAGCGAACGGGAUIIAG AIIACCC CG GCAGHCCA CGCAGH NA
-GAGCAAACAGGAI II 'AG Ai ACCC i G G!iAGUCCA--CGCCCiI-M
i l l 11 m i l i m i l l UBMIIlLim i " i ' ' " I MI m H I M MI I MI I I n I U I U I IIII
: CGC UGAGGil UCGA pAG G GGG !'AGCAAACAGGAIII'AG Al ACCC G Gl'AGI'CCA CACCG.I AA
i : CAC UGAGGC MMNA M G f G GGG GAGCAflftCflGGAU; AG Al'ACCC UG AGilCCA CGCCG'I AA
i :-CGC-IIGAGGA-GCGA flAGCG GGfl! GAGrGAACflGGAUhAG-AKACCC-nG Gi:AG!!CCA--CGCCGII-AA
:-CGC-IGAGGA-GCGA- AAGCG GGQ GAGCGIWCflGGA!!rAG-AHACCC-MG-GI'AGiiCCA--CGCCGll-AA
k : CGC UGAGGC GCGA AAG'-G CGfl GAGCGAACAGGAlillAG AUACCC HG GHAGIICCA CGCCGll AA
1 : CGC IIGAGGII GCGA AAGCGJGIJ i AGCGAACHGGAUUAG AUACCC HG GUAGHCIIA CGCCGU AA
t : CAC IJGAIIGC UCGA AAGUGUGGG i.'AllCAAArAGGAllHAG AUACCC IIG GIIAGIICCA CACAGII AA
c t-CGC-UGAGGU-GCHN-AAGCGUGGG-GAGCAflACAGGA'JllAG-AUACCC-UG-GUNGIICCACGCCGIJ-AA
V : CGC IIAAGGC GCGA AAGCAAGGG GAGCAAACAGGAUIIAG AIIACCC UG GIIAGIICCII IIGCCGII AA

Downloaded from http://bioinformatics.oxfordjournals.org/ at University of Michigan on September 19, 2015

t : CGC IIAAGGC GCGA AAGCAAGGG GAGCAAACAGGAUIIAG AIIACCC IIGGIIftGIICCII IIGCCGII AA

been written to provide the user with a full description of all

principles underlying DCSE, and of all parameters and functions
used. Pos:Qii66b Deaulfobacterlun udcuoiatun arteub.ali
Options: 41'iTiBffMHHffli Houi;scc(S):y Shouscc(u) Colour(c) Refresh B<
^ Sliou tuo ulndous on different parts of the allgnnent |
When DCSE starts up, the current directory is checked for c : G CtG U C C<G U A)G C - U G) [ECU - t C A G - U(G A GHS G - A C1G A
si: o oto o o o(o o o)D o - 0 0] to o o -tC G C - U<G A G>G C - G II IN A
the presence of a file containing defaults. It is recognized by B : G CtG C G C<G U A)G G - U Gl IG U U -tC A C - (KG A G}G U - G C1G A
I : G CtG U G C GC A>G G - C G] [G U U -tC G C - U(C A>U G C -A C1G A
its name. In this file, users can put their preferred default values s : G GIG C A C G(C A>G G - C Gl [G U U - [ C G C - U<C A)U G U - G C1N A
p : G CtG C A U(G C A>G G - U G] [G U(G - t C A C - U<C I E M II - G C1G A
for several parameters, such as the name of the alignment file u : G G(G(C)A C<G C A}GG ~ (Tfil tG U C - t C G C - UlN A G>G U - (A)CIN A
si: G GIG li G C(G C AJG G - C G l tG U C - t C G C - U#( A G>G C - G N]N A
and the reference file, or an alignment position to go to. Every rl: G AtG C G C(G U fl>G 6 - C G l tG G U - t C G C - t l A U)G C - G C1G A
c : G AtG U G C<G U D I G S - C Gl tG G U - t C G C - U<G A U>G C - A C1G A
parameter can have several defaults. Any default can be edited a : G GIG A A C<G C AJ6 G - C G | [G(U)C - t C G C - U1G A G>G U - U C1G A
G GIG C G C(G U A _ _ [ G u - [ C A C - U A G>G C - N N1N A
before it is selected. G AtG C U C<G U --[SIIC - I C G C - U(G A GW A - G C1G A
G GCG C U C<G U . - - I G U U -1C G C - U(G A G>G A - G C1G A
The first thing DCSE asks is the name of the alignment file. G GIG C G CtG U A
N S - C Gl tG U A -CC G C - tKG A G>G C -
G C1G
G C1G
A
A
G GIG C A OtG 11 A>G G - C Gl tG U U - t C G C - U<G A G)G U -
A list of all organisms present in this file is created. The G GIG A G CtG U A)G G - U G] IG A C - I C A C - U(G A U>G C -
U C1G
G C1N
A
N
G CtG C A C(G U MS G - C Gl tG U U - I C G C - U<G A GG U -
selection screen shown in Figure 2 is used to select which G GIG C G IKG U A)G G - C G l tG A A - t C 6 C - UOI A GIG C -
G C1G A
23 23-
organisms and positions will be loaded into memory. Organisms
can be selected individually or in block. A search function
allows users to find and select a specific organism. The current 9s:0Z676 Splrochaeta bajacalifoi
selection can be saved and retrieved later on when work on File PosiTion Prinary Sccundary ]JQ^^ Dluur^ Quit
| Options ncnu: s p l i t uindous, noue mid show secondAry structure,
the same set of organisms is resumed. si
rl
:M
:A
AAGGGHGCGCftSG-CS
AAGAGCGCGUAGG-C8
Gl
G
The two top lines on the screen show information and menus. 'c
a
;H
:A
AAGAGUGCSUAGG-CGGGUflS:__
AAGGSftnCGCAGG-CSrGUCCUU.
DCSE can disnlav the seauences in several modes, which are b :A
p :A AAGA
AAGGGCGCGUAGG-UG-GAUflUUI
'
: :A AAGG C GUCGC-GUCtl
b :A- -AAGG U- GGCftiV GUCfllffilSSACGC-UWOGC GCf.n
1 :A AAGG U--UGUHA-GUCC p-AU-UGA~CGC-UGAGGU-C(M
It :A AAGG fi GUUAA-GUCA AC-UGACAC-UGAUGC-UCGn
t :A -mOC ACUCAA-GUCAfalU-UGACGC-UGAGOJ-GCNN
fA :A-LAAGG.-_ IAG6AAA-GUUA ?-CC-UGACGC-UASfeGC-eCGn
. :A- 1 AAGGGCG PAAAGGUAH-GUUA MC-UGACGC-U(WG6C-GCGA-
i :A AAGG^CG GUGG-GCUGC-GUCq M-CC-UGfl--CGC-UBAGGC-GCGA-
r :A AAGG - -Gcps jicc-i(G(>-psc-j'""--
^.Mt;thij]i:ib<u:tf:riun s p . I 1 :A AAGG ^GUCillKAC-UGn --C6C-U
lP:A AAG- :-UGACAC-U
i :A
K Kijlosnxirfflns
inolMci 1 lu-i suis GCCCU6UGH-GUCU|[C-%X:-CG>1CCC-H,
i*.Uibrin n r o t c o l u t i c u s
9,Dcsulfobacter s p . 1
lO.Flextspira rapplni 1
ll.Hclicobacter ciiMedl
12.Lactobacillus agilis
13 Clostrldiun butyrlcun
pctroleopliiU Fig. 3. DCSE's editing screen in several display modes under DOS. The top
15-Blridobactcriim cumiiculi line shows information about the current position of the pointer, which is
bajacaliftirnirnxis
17 Si.irochoeU lltoralls displayed as a rectangular area with a grey background. The second line is
used for messages and menus. The next line displays help. The following are
sequence lines, preceeded by an abbreviation of the organism name: (a) DCSE's
editing screen in 'colour' mode. Secondary structure is not shown. Different
Fig. 2. The selection screen of DCSE under DOS. It shows the current alignment characters are displayed in different colours, seen here as shades of gray, (b)
file (arteub.ali), followed by the currently selected positions. The upper box DCSE in split-screen mode. The left and right windows display different parts
is a dialogue box which allows the user to find a certain organism or to select of an RNA sequence alignment with secondary structure information, (c) The
a set of organisms designated by a filename. The lower box displays a pan editing screen with the secondary structure indicated using colour. The same
of the organisms present in the alignment file. These can be scrolled up and helix as in (b) is displayed. Double-stranded areas are indicated by a different
down. Selected organisms are shown in inverse video. Options are shown on background colour, shown here as gray. Still other colours are used for bulges,
the screen and can be selected by pressing the highlighted characters. internal loops and non-standard basepairs.

738

illustrated in Figure 3. The next line is the help line. The other
lines contain sequences, preceded by an abbreviation of the
species name. The pointer is shown as a rectangular area in
inverse video or with a differently coloured background. Display
of the structure symbols can be switched off, in which case
the number of nucleotides visible on the screen is doubled.
Every character can be given a specific colour, which makes
it possible to display bases or amino acids, and/or secondary
structure elements in a different colour.
An extensive range of functions allows location of certain
features in the alignment. DCSE can go to a specified sequence
or alignment position. It can position the pointer on a specific
organism or helix. It can also find a certain sequence or its
complement, allowing for a specified maximum number of
differences.
For large alignments it takes a considerable amount of time
to save the entire file after alterations have been made, therefore
DCSE saves corrections directly into the file originally loaded.
It is thus possible just to save a few sequence lines while leaving
the rest unchanged. A global insertion or deletion will
necessarily need a complete rewrite of the alignment. A flexible
output routine enables the user to save any part of an old
alignment as a new partial one. This routine has an option to
remove all positions that do not have nucleotides in the selected
sequence lines. This way the new alignment can be compacted.
The output alignment can also be sliced into blocks of given
size in order to print it. The secondary structure signs can be
removed, and the names of the organisms abbreviated.
CONVERS is a program that complements DCSE. Several
options can be chosen from a central menu. All options are
menu driven as well. CONVERS currently has the possibility
to change files from IG Suite format (IntelliGenetics, Inc.,
Mountain View, CA) to EMBL format (European Molecular
Biology Laboratory, Heidelberg). It has an option to find and
extract sequence information from files obtained from EMBL,
and to convert them to the reference format (see Algorithms,
primary structure) used by DCSE. It can convert an alignment
or parts of the alignment to the format used by TREECON (Van
de Peer and De Wachter, 1993) for deriving phylogenetic trees.
It can also append sequences from a reference file to an
alignment. Information such as literature reference and
taxonomic position can be extracted in user-configurable formats
from a reference file. Other options include sorting of an
alignment or reference file, and appending different alignments
of the same length.
Discussion
Several sequence alignment editors, such as HOMED
(Stockwell and Petersen, 1987), MASE (Faulkner and Jurka,
1988), ALMA (Thirup and Larsen, 1990), SOMAP (Parry-
Smith and Attwood, 1991) and MALIGNED (Clark, 1992) have
Dedicated Comparative Sequence Editor
been described in the literature. Olsen's SEQEDT is available
from the Ribosomal Database Project (Olsen et al., 1991). The
GCG sequence analysis package (Genetics Computer Group,
Inc., Madison, WI) provides the multiple sequence editor
LINEUP. With the exception of SEQEDT and LINEUP, these
editors were originally developed to edit protein alignments,
though they can be used for nucleic acid alignments as well.
DCSE was developed from the start to deal with structural RNA
molecules, though it can also be used on protein alignments.
DCSE provides the tools to investigate higher-order structure
on a comparative basis. DCSE also has several other advantages
Downloaded from http://bioinformatics.oxfordjournals.org/ at University of Michigan on September 19, 2015
over other editors due to its special approach towards editing
and its portability.
While available editors are limited either to VMS or Unix
workstations, DCSE is available for different platforms. Its
function is not limited to workstations, and its easy approach
towards working on a part of an alignment makes handling of
large alignments possible, even on microcomputers. The
number of rows and columns visible on the screen is limited
only by what the system can handle, and not by DCSE. All
data files used by DCSE are essentially plain ASCII files in
a simple format. This makes data interchange easy. In contrast
to SEQEDT, DCSE does not incorporate functions to create
evolutionary trees. However, the auxiliary program CONVERS
can easily create data files which are used by TREECON (Van
de Peer and De Wachter, 1993) to produce trees. It would be
equally straightforward to create data for other phylogenetic
packages.
DCSE's strength lies in its unique supportunrivalled by the
other editorsfor secondary structure research. DCSE allows
incorporation of structural information into an alignment, so
that it can be used to refine the alignment. Incorporated RNA
structures can be copied, checked and displayed in several
modes (cf. Figure 3). Further tools for secondary structure
research are the possibility to search for the complement of a
sequence, the ability to split the screen, and a routine that helps
to locate possible helices by detection of compensating
substitutions. Protein structure information can also be
incorporated in a alignment. In fact any symbol not reserved
for RNA secondary structure can be used to indicate any
property tied to a certain residue. Giving hydrophobic and
hydrophilic residues a different colour can aid the detection of
structure elements. However, DCSE has no special support to
seek protein structures, such as site/signature searches.
DCSE offers a powerful and user-friendly environment to
create and maintain sequence alignments. It has been used
extensively in our research group for the maintenance of an
up-to-date alignment of small ribosomal subunit RNA sequences
(De Rijk et al., 1992). An alignment of large subunit ribosomal
RNA is also being created. A number of routines could still
be improved, such as the automatic alignment routine. A
criterion to choose whether a global insert is needed to
739
P.De Ryk and R.De Wachter

accommodate extra residues in a newly aligned sequence should

be considered. The different versions of DCSE for supported
platforms are available from the authors. They will also be made
available on-line through anonymous FTP on host
uiam3.uia.ac.be (143.169.8.1).

Acknowledgements
This research was supported in part by the Program on Interuniversity Poles
of Attraction (contract 23) of the Office for Science Policy Programming of
the Belgian State, and by the Fund for Collective Fundamental Research. It
was performed in the framework of the Institute for the Study of Biological
Evolution of the University of Antwerp. Peter De Rijk is a research assistant

Downloaded from http://bioinformatics.oxfordjournals.org/ at University of Michigan on September 19, 2015

of the NFWO.

References
Chan.S.C, Wong.A.K.C. and Chiu,D.K.Y. (1992) A survey of multiple
sequence comparison methods. Bull. Math. Bioi, 54, 563-698.
Clark,S.P. (1992) MALIGNED: a multiple sequence alignment editor. Compui.
Applic. Biosci., 8, 535-538.
Depiereux.E. and Feytmans.E. (1992) MATCH-BOX: a fundamentally new
algorithm for the simultaneous alignment of several protein sequences.
Comput. Applic. Biosci., 8, 501-509.
De Rijk,P., Neefs,J.M., Van de Peer.Y. and De Wachter.R. (1992) Compilation
of small ribosomal subunit RNA sequences. Nucleic Acids Res., 20,
2075-2089.
Faulkner.D.V. and Jurka.J. (1988) MASE: multiple aligned sequence editor.
Trends Biochem. Sci., 12, 279-280.
Olsen,G.J., Larsen.N. and Woese.C.R. (1991) The ribosomal RNA database
project. Nucleic Acids Res., 19, 2017-2021.
Parry-Smith.D.J. and Attwood.T.K. (1991) SOMAP: a novel interactive
approach to multiple proteinsequences alignment. Comput. Applic. Biosci.,
7, 233-235.
Rechid.R., Vingron.M. and Argos.P. (1989) A new interactive protein sequence
alignment program and comparison of its results with widely used algorithms.
Comput. Applic. Biosci., 5, 107-113.
Schuler.G.D., Altschul.S.F. and Lipman.D.J. (1991) A workbench of multiple
alignment construction and analysis. Proteins Struct. Fund. Genet., 9,
180-190.
Sellers,P.H. (1974) On the theory and computation of evolutionary distances.
SUM J. Appl. Math., 26, 787-793.
Spouge.J.L. (1991) Fast optimal alignment. Comput. Applic. Biosci. , 7 , 1 - 7 .
Stockwell.P.A. and Petersen.G.B. (1987) H O M E D : a homologous sequence
editor. Comput. Applic. Biosci., 3, 37-43..
Thirup.S. and Larsen.N.E. (1990) ALMA, an editor for large sequence
alignments. Proteins Struct. Funct. Genet., 7, 291-295.
Van de Peer.Y. and De Wachter.R. (1993) TREECON: a software package
for the construction and drawing of evolutionary trees. Comput. Applic.
Biosci., 9, 177-182.

Received on April 30, 1993; accepted on June 23, 1993

Circle No. 16 on Reader Enquiry Card

740