Transcript 174-16-Winter_9_2-Fe..

Final Exam:
Friday, 20 March 2015, 8:00 - 11:00 A.M.
In Spieth 2200 = same as lecture room
Emphasizes last 1/3 of course, but may
include carry-over things (e.g., residuals,
natural selection, multiple solutions)
Format will be similar to previous 2 exams
1
Your Teaching Assistant (Layla Hiramatsu)
will hold two extra office hours this week:
Thursday, 19 March, 2:00-4:00 P.M.
2378 Spieth Hall
You must come with questions to ask,
as she will not have any extra review
material prepared for you.
2
Your Professor (Dr. Theodore Garland, Jr.)
will also hold two office hours next week:
Thursday, 19 March, 9:30-11:30 A.M.
2366 Spieth Hall
You must come with questions to ask,
as he will not have any extra review
material prepared for you.
3
Please remember
to fill out course
evaluations!
4
Lecture 19:
Interspecific
Comparison and
Why Phylogeny
Matters
5
Four General Approaches to Studying Evolution:
1. Comparisons of Species (or populations)
= "The Comparative Method“
Shows what has happened in past evolution
2. Biology of Natural Populations
Shows present evolution in action
3. Selection Experiments
Shows, experimentally, what might happen
during future evolution
4. Comparison of Real Organisms with
Predictions of Theoretical Models
Shows how close selection can get to
producing optimal solutions
6
Three Sampling Schemes for Comparing Species:
1. Do two species differ in the predicted
direction? (not adequate)
2. Do multiple sets of species pairs differ in the
predicted direction? (acceptable)
3. Does the expected relationship hold across
many species? (the most general)
7
Three Sampling Schemes for Comparing Species:
1. Do two species differ in the predicted
direction? (not adequate)
Example: what are the respiratory adaptations to
living in burrows, where O2 levels can be low and
CO2 levels high?
Physiologists compared burrowing owls with
bobwhite quail.
Only the owl uses
burrows, and they
are about the
same body size.
8
But this comparison is ambiguous.
Any differences might be attributable to:
adaptations to a burrowing habit (owl)
carnivory (owl) vs. herbivory (quail)
(partial) domestication (quail)
nocturnal (owl, mostly) vs. diurnal (quail)
or any other aspect of "owlness" or "quailness"
(i.e., shared derived features
[synapomorphies] of the owl
clade or of the quail clade).
9
Garland, T., Jr., and S. C. Adolph. 1994.
Why not to do two-species comparative
studies: limitations on inferring adaptation.
Physiological Zoology 67:797-828.
Four limitations of two-species comparative
studies:
10
A. Independent variable is confounded with
species (lineage) membership.
The species may differ in many ways and for many
reasons, in addition to the one that is of interest
(e.g., whether they live in burrows).
11
B. Species almost always will differ.
1. The relationship between speciation per se (i.e.,
the evolution of reproductive isolation) and
phenotypic evolution is poorly understood,
but the process of speciation itself may result in
genetic differentiation that affects various
phenotypic traits.
2. The two species will (by definition) have
experienced little or no genetic exchange since
the time of their evolutionary divergence, and
so, at a minimum, they will have diverged
somewhat by random genetic drift alone.
12
3. They -- or their ancestors -- will have
experienced different environmental conditions
and hence different selective regimes, one axis
of which is why they were chosen for
comparison (e.g., whether they live in burrows).
Therefore, comparing two species is not like
comparing two groups of otherwise-identical
rats, one of whom received a drug and one of
whom received a placebo.
a. The appropriate null hypothesis is not no
difference, but rather some difference(s).
b. If difference is presumed to be likely, then you
have a 50% chance it will be in the predicted
direction, so Type I error rate is 50% not 5%.
13
C. Degrees of freedom for correlating trait
with environment = N - 2 = 0.
As soon as you go to three or more species, you
are protected from this problem.
2 species, d.f. = 0
3 species, d.f. = 1
Trait
Trait
Environment
Environment
14
D. Comparison of only two species would
not allow inference concerning which
state was ancestral.
If your definition of "Adaptation" is historical and
requires showing that it evolved from a different
state, then you cannot infer (e.g., by parsimony)
this with only two species.
A
O
O
O
O
15
Three Sampling Schemes for Comparing Species:
2. Do multiple sets of species pairs differ in the
predicted direction? (acceptable)
Example: Martin, T. E., P. R. Martin, C. R. Olson, B. J.
Heidinger, and J. J. Fontaine. 2000. Parental care and
clutch sizes in North and South American birds. Science
287:1482-1485.
Background: "The evolutionary causes of small clutch
sizes in tropical and Southern Hemisphere regions are
poorly understood."
Hypothesis: "Alexander Skutch proposed 50 years ago
that higher nest predation in the south constrains the rate
at which parent birds can deliver food to young and
thereby constrains clutch size by limiting the number of
young that parents can feed."
16
Assuming that each taxon is from a different
clade, then each pairwise comparison is
essentially an independent piece of information.
Any difference between two species within a
genus (or family) must have arisen since they last
shared a common ancestor, and that common
ancestor is shared to the exclusion of species in
other genera.
17
So, this method uses phylogenetic information in
a simple and incomplete way.
And, it does not make use of differences among
the species pairs.
North
South
Genus 1
North
South
Genus 2
North
South
Genus 3
18
8 pairs of species to be compared
19
The AZ
species
has a
larger
clutch in
each of the
8 cases.
paired ttest,
P < 0.001
20
The AZ
species
experience
a higher
predation
rate in 7 of
the 8
cases.
paired ttest,
P = 0.011
21
Extended data set with
additional species:
Overall Conclusion:
"Skutch's hypothesis
explains clutch size
variation within North and
South America.
However, neither
Skutch's hypothesis nor
two major alternatives
explain differences
between latitudes."
(Martin et al. 2000)
22
Three Sampling Schemes for Comparing Species:
3. Does the expected relationship hold across
many species? (the most general)
23
General Procedure for Comparative Study:
1. Measure several species (and/or populations)
for some phenotypic trait(s); calculate mean(s).
a) Should be raised in common environment
or "common garden," and really need at
least two generations
2. Characterize environmental features that
should indicate variation in the
"selective regime"
3. a) Relate interspecific phenotypic variation to
environmental variation -- evidence for
adaptation? (e.g., clutch size vs. predation rate)
b) Relate traits to each other -- elucidate
functional relations, allometry, trade-offs, etc.
24
Example: species in
several lineages of
mammals have
become secondarily
aquatic, to one
degree or another.
water shrew
water rat
muskrat
beaver
otter
mink
hippo
seal
sea lion
manatee
whales
25
Water Shrew
Water Rat
26
Muskrat
27
28
Otters of various types, including the Sea Otter
(Enhydra lutra)
29
Mink
30
31
32
The mammals exhibit a gradation in how aquatic
they are (some barely, some completely).
Not all are members of the same evolutionary
lineage (clade), so this habit has evolved multiple
times.
Allows us to formulate and then test simple
adaptive hypotheses.
The more aquatic the animal, the more
hemoglobin it will have in its blood.
33
Natural
& Sexual
Selection
Act
On
Behavior
This hypothetical
example is
essentially trying to
relate these levels.
Morphology, DeterPhysiology,
Biochemistry mine
Organismal
Performance
Abilities
34
[Hemoglobin] in Blood
3
Hypothetical Example
2
N = 50
1
d.f. = 48
0
R = 0.517
-1
1-tailed
P < 0.0001
-2
-3
-3
-2
-1
0
1
2
3
How Aquatic the Species Is
35
Real Examples,
Some Shown in
Previous Lectures
(interspecific
comparisons are very
common in ecological
and evolutonary
physiology
36
Natural
& Sexual
Selection
Act
On
Morphology, DeterPhysiology,
Biochemistry mine
Behavior
Organismal
Performance
Abilities
37
Amphibians that
live nearer to the
equator have
higher thermal
tolerances.
"These data are consistent
with the thesis that
congeneric and conspecific
populations in different
habitats have partially
independent evolutionary
pathways and that where
populations have been
displaced from a less to a
more uniform environment,
characters which are no
longer maintained by
selection pressure are lost."
(p. 100)
Snyder, G. K., and Weathers, W. W.
1975. Temperature adaptations in
amphibians. American Naturalist
109:93-101.
38
Natural
& Sexual
Selection
Act
On
Morphology, DeterPhysiology,
Biochemistry mine
Behavior
Organismal
Performance
Abilities
39
Maximum isometric
twitch tension of lizard
muscles is positively
correlated with PBT
across species.
"These data are interpreted as indicating
that divergence in the thermal
adjustments for contractility of muscular
tissues are most directly associated with
the temperatures characteristically
maintained through the behavior of the
species rather than to the general
geographical (climatic) distribution or
phylogenetic position of the species."
(p. 1)
40
More Real Examples ...
41
Do canids (endurance
athletes) have larger hearts
than felids (sprinters)?
42
43
Do desert animals have greater urineconcentrating ability?
141 species or populations of rodents
Habitat coded as Aquatic, Mesic, Semi-desert,
Desert
Al-kahtani, M. A., C. Zuleta, E. Caviedes-Vidal, and T. Garland, Jr. 2004. Kidney mass and relative
medullary thickness of rodents in relation to habitat, body size, and phylogeny.
Physiological and Biochemical Zoology 77:346-365. (plus online Appendix B).
44
Yes, they do.
Also note the negative
allometric relationshpp.
45
Do desert animals have reduced basal
metabolic rates?
Deserts are generally hot and dry, and have low
primary productivity.
Reduced BMR would reduce needs for food and
water, and could reduce chances of overheating.
31 populations/species of deer mice
(Peromyscus) were studied.
Mass-corrected BMR was correlated with
various characteristics of the environment
(temperature, precipitation) at or near their point
of capture.
46
47
Residual log B.M.R. (ml O2 /g x h)
0.175
31 Populations of Peromyscus
0.125
0.075
0.025
-0.025
-0.075
-0.125
r2 = 0.27,
2-tailed P < 0.005
-0.175
11 13 15 17 19 21 23 25 27 29 31
Mean July Temperature (oC)
48
What factors affect size of the testes in
primates?
Chimp
Brain
Chimp
Testicle
Main source is:
Harcourt, A. H., P. H. Harvey, S. G. Larson, and R. V. Short. 1981.
Testis weight, body weight and breeding system in primates. Nature 293:55-57.
See also: Harvey and Pagel (1991) figure 1.3
Harvey, P. H., and M. D. Pagel. 1991. The comparative method in evolutionary
biology. Oxford University Press, Oxford. 239 pp.
http://www.atomicnerds.com/wp-content/uploads/2010/08/Picture1-5.png
49
Primate Species: Allometry
Testes Mass (g)
100
10
1
0.1
1
10
Body Mass (kg)
100
50
Primate Species: "Residual Variation"
What accounts for the
huge (almost 10-fold)
variation at a given
body size?!?!
Testes Mass (g)
100
10
1
0.1
1
10
Body Mass (kg)
100
51
Primate Species: Sexual Selection
Testes Mass (g)
100
Females likely
to copulate
with > 1
partner per
oestrus
10
Females
mate with
only 1 male
1
0.1
1
10
Body Mass (kg)
100
52
How do Human Beings Compare?
Females likely to
copulate with > 1
partner per oestrus
Testes Mass (g)
100
Homo
sapiens
10
Planet of the Apes (1968)
Females
mate with
only 1 male
1
0.1
1
10
Body Mass (kg)
100
53
The Basic Statistical
Problem Posed by
Comparative Data:
related species
tend to resemble
each other ...
54
For example, all cats are recognizable as such:
http://yourlifestyledestination.com/wp-content/uploads/2014/02/wildcats.jpg
http://wildcatfund.org/catmontage.jpg
55
… so they do not necessarily
represent independent pieces of
information (evidence) when
testing for relationships between
traits, differences among groups
of species, adaptive hypotheses,
effects of sexual selection, etc.
56
Physiologist
sees this
Close
Relatives
Tend to
Resemble
Each Other
57
What
Conventional
Statistical
Methods
Assume
What
Evolution
Provides
58
Statistical Consequences of Ignoring
Phylogenetic Relatedness
Type I error rates will be inflated if species
are treated as independent:
null hypothesis will be rejected too often.
Power (ability to detect relationships)
will be affected.
Estimates of correlations, slopes (e.g., in
studies of allometry) or group
differences will be inaccurate.
59
Phylogenetically Based Analytical
Methods Allow One To:
1. avoid statistical problems caused by
phylogenetic non-independence;
2. learn more from the data, e.g.:
a. Infer ancestral states
b. Compare rates of evolution
(phylogenetic lability) across
lineages or across traits
c. Test for "outlier" species
d. Discover new patterns in the data
60
In addition, consideration of
phylogenetic information allows one
to make better choices as to which
species should be compared!
61
Hypothetical Example:
Do animals that live at high altitude have
more hemoglobin in their blood?
62
Hemoglobin (g/100 ml)
The Scattergram:
23
21
19
17
15
13
11
9
7
-500 0 500
1500
3000
4500 5500
6000
1500 2500
3500 4500
6500
Altitude
Altitude (m)
(m)
63
Hemoglobin (g/100 ml)
Worst-Case Scenario:
23
D
21
Another
Lineage
19
17
13
One
Lineage
11
9
E
B
15
F
C
A
7
-500 0 500
1500
3000
4500 5500
6000
1500 2500
3500 4500
6500
Altitude
Altitude (m)
(m)
64
Phylogenetic Pseudoreplication
L
L
Red = low
altitude
L
H
H
H
Blue = high
altitude
65
Conventional Statistics Assume:
L
H
Red = low
altitude
L
L
H
H
Blue = high
altitude
66
Best-Case Scenario:
H
L
Red = low
altitude
H
L
H
L
Blue = high
altitude
67
But even if species are chosen
poorly, phylogenetically based
statistical methods will allow you
to obtain the "correct" answer.
They can rescue you from poor
experimental design ...
68
Phylogenetically Independent
Contrasts (Felsenstein, 1985)
1. first fully phylogenetic method
2. widely applicable (correlation, PCA,
regression, ANCOVA)
3. has been extended in several ways
(incomplete phylogenetic information,
including arbitrary branch lengths,
sensitivity to deviations from Brownian
motion, estimates of ancestral values)
4. mathematically equivalent to GLS
69
Branch Lengths
in Myr
X 8
A 9
Body
B 7 Mass
(g)
C 2
Identify and Compute Independent Contrasts
Compute square roots of sums of
(corrected) branch lengths = S.D.
Standardized
Contrast Value S.D.
Contrast
A-B
X-C
2
6
2
3
1
2
70
Branch Lengths
in Myr
X 5
A 4
Bill
B 6 Length
(cm)
C 3
Identify and Compute Independent Contrasts
Compute square roots of sums of
(corrected) branch lengths = S.D.
Standardized
Contrast Value S.D.
Contrast
A-B
X-C
-2
2
2
3
-1
0.667
71
r2 = 0.30
Body Mass (g)
Bill Length
Bill Length (cm)
It can make a big difference!
r2 = 0.02
Body Mass
Note that degrees of freedom are the same!
72
Felsenstein, 1985, p. 14:
"Comparative biologists may understandably
feel frustrated upon being told that they need to
know the phylogenies of their groups in great
detail, when this is not something they had much
interest in knowing.
Nevertheless, efforts to cope with the effects of
the phylogeny will have to be made.
Phylogenies are fundamental to comparative
biology; there is no doing it without taking them
into account."
Felsenstein, J. 1985. Phylogenies and the comparative method. American Naturalist 125:1-15.
73
Some Real Examples in
which a Conventional
and a Phylogenetically
Informed Statistical
Analysis Give Very
Different Results
74
Estimating Allometric Relationships
75
Basal metabolic rates of
254 species of birds:
Conventional slope and 95% C.I.
0.652 < 0.670 < 0.687
Independent Contrasts
0.687 < 0.722 < 0.757
Branch lengths = DNA-0.2
Reynolds, P. S., and R. M. Lee, III. 1996. Phylogenetic analysis of avian energetics:
passerines and nonpasserines do not differ. Am. Nat. 147:735-759.
76
Extra Slides Follow
http://www.transmittingscience.org/courses/evol/pcm/
This lecture was about 10 minutes too short in 2012.
For next time, add a few examples in, which makes a big difference.
However, was only about 5 minutes short in 2013.
For 2014, I deleted some things to reduce overlap with 105 lecture,
I replaced Moller with Martin, I added Estimating Allometric
Relationships in Birds, and the time was about 5 min short.
For 2015 it was fine, but need to swap more examples for 2016.
This lecture is still almost identical to
105_2013_10_31_Comparative_Methods_1.ppt
slides 34-104
77
log10 Home Range Area (km2)
Conventional P for diet < 0.001
Independent Contrasts P = 0.15
3
Carnivora
2
Eat
Herbivores
1
0
ungulates
-1
Eat Plants
-2
0
1
2
3
log10 Body Mass (kg)
78
ungulates
Diet and clade are
completely
confounded, and the
comparison boils
down to one
contrast, so have low
statistical power.
79
Natural
& Sexual
Selection
Act
On
Morphology, DeterPhysiology,
Biochemistry mine
Behavior
Organismal
Performance
Abilities
80
This is used in 105 Evolution
Time required for exposure to 37oC to inactivate
myofibrillar ATPase is positively correlated with
thermal environment across species of fish.
Johnston, I. A., and Walesby, N. J. 1977. Molecular mechanisms of temperature adaptation in fish
myofibrillar adenosine triphosphatases. J. Comp. Physiol. 119:195-206.
81
Natural
& Sexual
Selection
Act
On
Morphology, DeterPhysiology,
Biochemistry mine
Are two behavioral
traits correlated?
Behavior
Organismal
Performance
Abilities
82
10 Strains of Drosophila
Compared
r = -0.11;
this estimates the broad-sense
genetic correlation
Durcan, M. J., and D. W. Fulker. 1983. A comparative study of two Drosophila learning tasks. Behavior Genetics 13:179-190.
83
Amphibians
that live
nearer to the
equator have
higher thermal
tolerances
(Snyder and
Weathers,
1969).
84
Why generic
averages?
85
Why generic
averages?
86
Natural
& Sexual
Selection
Act
On
Morphology, DeterPhysiology,
Biochemistry mine
Behavior
Organismal
Performance
Abilities
87
Notes from 13 March 2008:
It is a bit too long. Possibly cut:
one primate generic averages example?
cut both of them?
Shrink 2-species?
88
89
See also:
Slidescm.doc (includes SEMICVM5.doc,
SEMIUPS7.doc, SEMOZ93.doc, )
SEMUBC94.doc
SEMEVO93.doc
90
Reminder!
Paper critique 3 due on Thursday, 15 March
Must be a paper from the journal Evolution or
Physiological and Biochemical Zoology, published
in 2006 or 2007.
Must be a multi-species (> 2) comparative study
that involves at least some physiology (or
biochemistry or morphology) and must offer some
ecological or evolutionary context for interpreting
the results.
You must turn in a photocopy of the entire paper
along with your critique.
91
Reminder!
Sample exam questions (3 points extra
credit) are also due on Thursday, 15 March
92
New seminar course for Fall 2012:
BIOL 190 – Research Careers in the Biological Sciences
• Professional-development training for undergrads interested in
research careers
• 1 credit
• Will meet once per week for 2 hours; day and time TBA
• Topics will include:
– Choosing a graduate program
– Applying to graduate school
– Conducting undergraduate research
– Interview skills
– The ins and outs of graduate school
– Types of research careers, and what to expect from them
• For additional information contact Prof. Wendy Saltzman
([email protected])
93