Running head: Fast method to combine distance matrices
48 pages
English

Découvre YouScribe en t'inscrivant gratuitement

Je m'inscris

Running head: Fast method to combine distance matrices

-

Découvre YouScribe en t'inscrivant gratuitement

Je m'inscris
Obtenez un accès à la bibliothèque pour le consulter en ligne
En savoir plus
48 pages
English
Obtenez un accès à la bibliothèque pour le consulter en ligne
En savoir plus

Description

Running head: Fast method to combine distance matrices SDM: a Fast Distance-based Approach for (Super)Tree Building in Phylogenomics Alexis Criscuolo 1,2 , Vincent Berry 2 , Emmanuel J. P. Douzery 1 and Olivier Gascuel 2,? 1 Groupe Phylogénie Moléculaire. ISEM, Université Montpellier 2, CC 064, 34095 Montpellier Cedex 05, FRANCE. 2 Equipe Méthodes et Algorithmes pour la Bioinformatique. LIRMM (CNRS, Université Montpellier 2), 161 rue Ada, 34392 Montpellier Cedex 05, FRANCE. ? Corresponding author: Tel. (33 or 0 from France) 4 67 41 85 47 Fax. (33 or 0 from France) 4 67 41 85 00 Keywords: phylogenomics, evolutionary distances, distance method, supertree, superma- trix, MRP, total evidence 1

  • large phylogenies

  • methods

  • such genome-sized

  • inferred using

  • single tree topological

  • large gene


Sujets

Informations

Publié par
Nombre de lectures 31
Langue English

Extrait

ast
to
dist
1,2 2 1 2,∗
1
oup
ISEM,
2
FRANCE.

3o
r0
41
85
00
total
Building
in
o
la
lier
Ph
e
ylogenomics
erma-
Alexis
a
Criscuolo
rtree,
(Sup
M
for
trix,
h
da,
,
ead:
Vincen
gascuel@lirmm.fr
t
F
Berry
ax.
Approac
des
,
p
Emman
LIRMM
uel
el
J.
rue
P
Montp
.
dex
Douzery
Corresp
Distance-based
http://www.lirmm.fr/~gascu
and
el.
Olivier
0
Gascuel
4
ast
sup
F
3
a
M?tho
SDM:
et
,
lgorithmes
trices
our
e
Bioinformatique.
Phylo
(CNRS,
g?nie
Montp
Mol?
lier
culair
161
e.
A
ma
34392
Universit?
el
Montp
Ce
el
05,
lier
h
2,
onding
evidence
uthor:
combine
el
F
T
4
(33
method
r
er)T
from
ords:
rance)
ylogenomics,
sup
olutionary
e
distance
F
F
(
Equip
CC
064,
from
34095
rance)
Montp
d,
Running
el
lier
Keyw
RP
ph
dex
ev
05,
distances,
FRANCE.
metho
A
1
67
47
85
41
67
2),
Universit?
Ce
Gr
ree
anceof
be
ork,
more
Ho
ev
es
poo
olv
el
Sup
SDM
erage
to
be
as
has
whic
a aO(n k ) n
be k
be a< 2
distance
be
eral
to
hes
es
when
taxa
the
also
can
be
eu
SDM
olv
s
are
useful
for
ject
The
(
exploratory
ds,
s
genes
tudies
i
and
b
building
taxa,
a
are
s
t
tarting
w
tree
minimization
to
of
b
system
e
genome-sized
rened
indicate
b
quic
y
7
a
w
distance-based
standard
p
the
o
w
w
t
erful
least-squares
maxim
onstrain
um
ecause
lik
linear
eliho
practical
o
require
d
um
(ML)
um
approac
,
h.
e
framew
ylogenies
w
fast
ev
requiring
er,
w
estimating
a
the
t
olutionary
with
distances
d
directly
ha
from
W
concatenated
this
genes
equiv
conrm
o
In
f
examined.
s
r
linear
top
This
ological
unique
signal
is
as
resolving
genes
enes
and
sparse,
taxa
g
e
ast
at
time,
dieren
the
t
r
rates.
that
W
r
e
homologous
prop
h
ose
sim
a
mammals.
no
obtained.
v
of
of
osed,
metho
studies
d
approac
,
o
y
sim
s
h
er
SDM
Distance
elev
Matrix
lternativ
(
t
r
Represen
),
arsimon
whic
)
d
notably
ylogenomic
h
follo
t
ws
e
the
v
same
conclusions.
line
in
as
problem
A
s
v
alen
um
t
Consensus
the
Sup
o
ertree
a
(A
criterion
heterogeneit
ub
Lap
h
o
c
in
ts.
ate
problem
and
a
Cucumel,
solution
1997)
h
and
obtained
com
y
bines
a
the
system.
r
this
olutionary
is
distances
its
obtained
resolution
trong
b
s
metho
h
f
e
data
in
these
to
is
a
n
single
h
distance
of
sup
Suc
ermatrix
n
n
genes.
large
of
analyzed
and
using
sets
a
esults
standard
r
distance-based
the
algorithm.
sup
SDM
rmatrix
deforms
from
the
kly
source
tal
matrices,
uses
without
placen
mo
prop
difying
5
their
exploratory
top
ph
ological
accurate
message,
build
to
f
bring
computing
them
Using
ypically
ulations,
close
to
v
o
p
48
o
is
ssible
r
a
an
eac
a
t
e
the
o
to
he
an
Matrix
t
tation
tree
P
an
y
approac
es
whic
metho
b
,
reduces
ulation
computing
studies
and
of
the
dieren
ogical
genes
ccuracy
v
W
lo
s
o
metho
erlap.
analyze
e
d
sho
of
v
et
h
other;
used
these
build
deformed
excellen
matrices
starting
are
for
t
ML
hen
h,
a
h
v
oth
eraged
the
to
that
obtain
increases
the
top
distance
a
sup
.
Abstract
ermatrix.
W
e
e
to
sho
the
w
ataset
t
Gatesy
hat
(2002)
to
2
al.
time
SDM
that
sets
MRP
that
time.
vier
hea
more
from
SDM
Sev
to
ws
allo
whic
matrices
the
where
requires
As
to
as
gene
eac
from
ev
te
CS,
named
ev
giv
this
large
aim
PhPh
yp
uge
al.,
ell
of
the
some
ha
be
can
be

(or
genes
obtain
con > 90%
Drisk
to
suc
but
seems
be
vide
and
al.,
2002).
MrBa
be
and
en
Ho
ev
f
sequenced
genes
from
ts,
stan-
w
probabilistic
ithin
Again,
a
some
h
e
is
with
v
can
ariet
ermatrix
y
As
of
umerous
organisms
ph
(Daubin
among
et
h
al.,
Genes
2002;
f
Gatesy
1
et
substitution
genes,
is
2002;
y
Eisen
re
and
usually
F
(e.g.
raser,
2004).
2003;
s
esian)
rmatrices
of
c
et
trend
al.,
b
2
e
004;
under
Philipp
of
e
m
e
problematic
t
e
al.,
s
2004;
s
D
w
evulder
analyzed
et
ph
y
econstruction
2005;
genes
Philipp
f
e
sup
et
tain
al.,
c
2005).
t
One
don
sets
v
large
metho
main
nets
diculties
sup
in
genes
Ba
dened
ylogenomics
b
is
that
(e.g.
m
sophisticated
and
most
to
d
data
s
al.,
with
mo
required
t
to
heterogeneit
pro
and
cess
v
the
d
large
b
collections
Y
of
Pupk
taxa
p
and
metho
genes.
tly
Missing
Huelsen
o
up
c
ecially
sp
haracters,
another
h
y
dicult
y
using
with
dard
suc
ositions).
h
r
datasets,
algorithms.
as
some
from
a
genes
missing
or
or
sp
taxa,
ecies
ermatrices
are
increased
less
n
represen
missing
ted
haracters
ssue,
he
databases.
in
Numerous
p
approac
al.,
i
co
built
arious
v
ylogenetic
e
d
are
among
e
analyze
n
h
prop
e
osed
(or
ain
or
deal
vulnerable
m
missing
this
haracters,
problem
the
(Bininda-Emonds,
e
2004);
to
a
not
ylogenies
uc
h
aected
classied
still
in
lar
is
trees
three
sparse
main
(Philipp
categories
dels
time
2004).
hmidt,
ev
2003,
e
c
dieren
computing
constrain
7):
opu-
whereb
y
The
rates
lo
o
w-lev
e
el
olutionary
ylogenomics,
o
total
es
evidence)
also
metho
e
ds
(
concatenate
ang,
er,
996;
approac
o
c
t
quan
a
2001)
probabilistic
vide
d
a
(e.g.
to
y
v
,
t
oduction
dicult
k
,
Ronquist,
y
pro-
dieren
w
alignmen
ys
t,
circum
curren
also
called
this
to
y
Intr
b
a
t
single
3
ylogen
hes.
for
wing
allo
es
olv
et
accurate
pro
ones
to
less
more
are
used
The
et

  • Univers Univers
  • Ebooks Ebooks
  • Livres audio Livres audio
  • Presse Presse
  • Podcasts Podcasts
  • BD BD
  • Documents Documents