Relative Rate Test

1
Relative Rate Test

• Marta Riutort

2
The molecular clock

• Is not perfect
• the nucleotid substitution rate, from a strict
  point of view, is never the same for all the
  lineages we compare
• This heterogeneity of rates generally is not
  going to have consequences in the inference of
  phylogenies, however

3
Long branch attraction (LBA)

• When the length of the branches or the
  substitution rates are extremely unequal, there
  is a violation of the assumptions made by
  inference methods that can produce the
  inconsistency of any of those methods
• We are entering the Felsenstein Zone

4
Felsenstein zone
5
Felsenstein zone
Filogenia real
Filogenia estimada
6
(No Transcript)
7
Equinodermo
DEUTEROSTOMADOS
Cefalocordado
Artrópodo
Nemertino
Molusco
Anélido
Políclado
Macrostómido
Nemertodermátido
PROTOSTOMADOS
94
Proseriado
Catenúlido
Rhabdocelo
100
Tríclado
Tríclado
Nemátodo-1
Nemátodo-2
Cnidario
Placozoo
0.10
Ctenóforo
Coanoflagelado
8
equinodermo-1
DEUTEROSTOMADOS
hemicordado-1
cordado-1
cordado-2
artrópodo-1
artrópodo-2
nematomorfo-1
nematodo-1
Ecdisozoos
priapúlido-1
kinorrinco-1
nemertino-1
sipuncúlido-1
anélido-1
Lofoforados Eutrocozoos
molusco-1
PROTOSTOMADOS
anélido-2
lofoforado-1
molusco-2
pogonóforo-1
anélido-3
lecitoepiteliado
proseriado
nemertodermátido
prolecitoforado
Platelmintos
catenulado
catenulad
macrostómido
macrostómido
políclado
ctenóforo
porífero
cnidario
9
fast-clock organisms

• These organisms with long branches are called
  fast-clock
• They really acumulate substitutions faster than
  the rest of organisms (their rate of substitution
  is higher)
• Some authors have proposed various hypothesis to
  try to explain this phenomenon
• higher metabolic rate, shor generation time,
  differences in the number of replications of DNA
  inthe germinal line, deficiences in DNA repair,
  mutagens,

10
Solutions?

• Use methods less sensible to this type of
  inconsistency (ML?)
• If it is possible, eliminate long branches
• eliminate the fast-clock organism
• substitute by another of the same group that is
  not fast-clock
• increase the number of organisms of that group

11
Solutions?

• We first need to know if we really have a
  fast-clock organism
• Relative Rate Test
• Sarich and Wilson, 1973 for proteins
• Wu and Li (1985) and Li and Tanimura (1987)
  extend it to nucleotides

12
Relative Rate Test

• Uses 3 species A, B and one outgroup C
• Tests if A and B have the same rate of
  substitution since their split
• dAO dBO
• dAC dBC
• d dAC - dBC 0

13
Relative Rate Test

• This method is time independent
• We have to be sure about the phylogeny

14
Relative Rate Test

• Our nule hypothesis is
• d dAC - dBC 0
• It is assumed that the number of nucleotide
  substitutions follows a Poisson,
• then we can use the standarized normal
  distribution to test if the number of
  substituions in the 2 lieages is the same

15
Relative Rate Test

• d dAC - dBC 0
• d Var(d)
• Var(d) Var(dAC) Var(dBC) - 2 Cov (dAC,dBC)
• Si
• d gt 1,96 Var(d) test is significative at 5
  level
• d gt 2,96 Var(d) test is significative at 1
  level

16

• Si
• d gt 1,96 Var(d) el test es significativo al 5
  
• d gt 2,96 Var(d) el test es significativo al 1

• Hipótesis nulad dAC - dBC 0

0
17
Equinodermo
DEUTEROSTOMADOS
Cefalocordado
Artrópodo
Nemertino
Molusco
Anélido
Políclado
Macrostómido
Nemertodermátido
PROTOSTOMADOS
94
Proseriado
Catenúlido
Rhabdocelo
100
Tríclado
Tríclado
Nemátodo-1
Nemátodo-2
Cnidario
Placozoo
0.10
Ctenóforo
Coanoflagelado
18
Relative Rate Test