Spider Evolution: Mutation, Selection, Drift

Introduction. As we learned earlier in the semester, spiders are a diverse group of

invertebrate predators that have radiated to fill many niches. But what evolutionary forces
have impacted this diversification? And is it possible to trace the events that occurred through
evolutionary time? Phylogenetic trees can help us organize evolutionary relationships among
related groups, but it’s important to understand the techniques used to construct these trees,
in order to assess their accuracy.

Rationale. There are two major goals of this laboratory exercise:

• You will learn to construct, interpret, and compare phylogenetic trees, using
morphological, behavioral, and molecular information from various spider taxa. You
will use this information to make inferences about, first, the evolutionary history of
spiders and, then, of spider silk.
• You will gain an understanding of how different mechanisms of evolution may have
contributed to spider diversity. Specifically, you will explore how mutations, selection
and genetic drift can shape variation in a lineage.

Part 1 Tree thinking: Using morphological, behavioral, and molecular data to explore the
evolutionary history of Spider families

Phylogenies are visual representations of evolutionary relationships, constructed using different

traits or character states. Character states can be morphological, behavioral, genetic, or
genomic.

Review the slides provided on Canvas that contain background information on several
representative spider families, including those used in the Spider Locomotion lab. First, we will
generate a tree based on morphological and behavioral traits.

Morphological and Behavioral Traits

From Canvas, download three files:

• Spider Phylogeny Exercise Info (PowerPoint file)
• Table of Derived Spider Characters (Word file)
• Spider Click-and-Drag Phylogenetic Tree (PowerPoint file)

Question 1. Open the Spider Phylogeny Exercise Info file and read through the slides,
paying attention to the descriptive notes as well as the photos of the spiders.

Use the information provided in the Info file to fill out the Table of Derived Spider Character
States, below.

1
Table 1. Derived Spider Character States

Tarantula Orb Net Wolf House Cellar

Jumping Nursery
outgroup weavers casting

Sticky Silk n/a

Independence
from web
n/a
(thick legs)

Elongate
Body n/a

Reduced silk
n/a
production
Parental Care
of Spiderlings n/a

Mouthparts
Inward

Mark an “X” in the cell when a spider family has the trait; leave cell blank if the spider family does not exhibit that
trait.

Question 2. Use the completed Table 1 to help you organize the spider families into a
phylogenetic tree using the Spider Click-and-Drag Phylogenetic Tree file from Canvas; this
PowerPoint file allows you to click on a spider photo and drag the photo to its appropriate
location on the tree. Map the above six traits onto the tree and rearrange the spider photos
to show the resulting phylogeny. When you are done constructing your tree, take a screenshot
of the phylogeny and insert it into the Assignment document in the appropriate place.

Question 3. Flower spiders are a group of spiders that lack sticky silk, are web-dependent, are
round-bodied with reduced silk production, have no parental care of spiderlings and exhibit
inward-facing mouthparts. Adding flower spiders to our phylogeny, using these derived traits,
would place flower spiders (family Thomisidae) as a sister to Cellar spiders; however, flower
spiders are actually believed to be more closely related to nursery spiders.
What does this suggest about the importance of selecting characters?
What might we do to resolve the ambiguity in this relationship?

2
Molecular Data

How do researchers generate trees from molecular data? Let’s focus on a small subset of data
from research on spider silk. Major ampullate spidroin 1 is one of the proteins produced by
spiders to make silk. Once released by the spider, repetitive protein filaments aggregate to
form strong silk. However, while still stored internally before release, these silk proteins are
maintained in liquid non-aggregated form with aggregation being prevented by conserved
non-repetitive protein regions, called N-terminal domains.

Researchers have compared the genomic sequence data for the N-terminal domain of Major
ampullate spidroin 1. Below is some alignment data showing several variable sites of amino
acids (represented with capital letters) when examining this portion of the genomes of four
spider species.

Position .............................................. 1 2 3 4 5 6 7 8 9

Diguetia canities ................................ - Q K E – M – – I

Latrodectus hesperus......................... - E K E – M – – I
Agelenopsis aperta ............................ - E Q A – I – – I
Nephila clavata .................................. - E Q Q – L – – L

Question 4. Fill in the table below with the variable sites and their sequences for each taxon.

Table 2. Variable sites in N-terminal domain of Major ampullate spidroin 1 in spiders

Diguetia Latrodectus Agelenopsis Nephila

canities hesperus aperta clavata
Position (Diguetidae) (Theridiidae) (Agelenidae) (Araneidae)
2
3
4
6
9

Question 5A. Use the information in Table 2 to complete the tree below. Use D. canities as
the outgroup. Note where each site difference occurs on the tree.

3
 D. canities
(Outgroup)

Question 5B. Below is sequence data for another spider species, Argiope bruennichi. Using
this sequence data, fill in the blank column in Table 2 for A. bruennichi. Now determine its
appropriate placement on the tree:
• Where is the best fit for Argiope bruennichi? (You can describe the location rather than
drawing it.)
• Explain why you chose this location.

Argiope bruennichi ............................ - E Q Q - L - - I

Part 2 Tree thinking: Using molecular tools to explore the evolutionary history of Spider silk

Spider silk

Spiders are a diverse group of predators with a wide range of prey-capture strategies, but the
most familiar strategies involve the use of silk (Fig. 1 Diversity of spider prey acquisition and
web architecture). Silk structure and web design are complex and closely linked to spider

4
diversity. A large body of research supports an early evolutionary origin for a type of silk known
as dragline silk. From there, the story is complicated and fundamental questions about the
evolution of spider webs are an active field of research.

Figure 1. Diversity of spider prey acquisition and web architecture. (Vollrath 2007)

To help untangle these questions, we will focus on one common style of web, orb webs. “Orb
web” is a descriptor used for a variety of architectures, but generally describes a web with
some radial or spiral aspect to its structure. Orb webs can be further described by the type of
silk used in their construction: Some use a non-adhesive silk (cribellate), while others employ a
sticky, adhesive silk (ecribellate). (This is a simplification of the many types of silk, but is
sufficient for our purposes.)

For decades, the prevailing hypothesis suggested that orb webs evolved only once, with sticky
silk orb webs deriving more recently from non-sticky silk-producing ancestors.

Question 6. Draw a tree demonstrating this hypothesis. Map the traits onto the tree. Recall
that dragline silk is the ancestral state. (Include a screenshot or photo of your tree.)

5
In 2009, Blackledge and colleagues generated the tree in Figure 2, using a combination of
molecular, morphological, and behavioral characters. Note that they found that sticky silk orb
web spiders further diversified, with some groups evolving specialized glues, or different silk
proteins, or independence from webs.

Figure 2. Non-orb web and orb web lineages. The bracket at the top includes all non-orb web
relatives. Open squares indicate loss of prey capture webs. Pie charts at left relate to statistical

6
analyses that are not necessary for our purposes; focus on the branching structure. (Blackledge
et al. 2009)

Question 7. According to the tree provided by Blackledge et al., how many times did orb web
architecture evolve? Is this tree consistent with your earlier tree from Q6?

As newer molecular tools became available, more researchers explored orb web evolution. In
2018, researchers used a larger molecular data set along with morphological data to generate
a new tree (See Fig. 3 Tree from Fernández 2018). The details are not important, but the red
circles indicate orb webs.

Figure 3. Ancestral state reconstruction using phylogenomic data set. (Fernández 2018)

7
Question 8. How many times did orb webs evolve based on Fernandez’s tree? Mark on the
tree where it occurred or include a screenshot.

Researchers continue to explore the evolutionary history of orb webs and find conflicting
results. In 2019, Sharff et al. used an even larger molecular data set than previous researchers
to generate a phylogeny. They constructed a phylogeny based on genetic characters and a
phylogeny based on non-molecular characters. Unfortunately, many of the genetic-based
clades could not be supported with non-molecular data, and vice versa. So even the same
research team was unable to reconcile the phylogenetic trees constructed with different lines
of evidence! This highlights one of the current challenges in resolving the ultimate spider
phylogenetic tree: a mismatch between ecological, biogeographical, morphological and
molecular data sets.

Discuss: What additional information might be useful in building a better tree?

Part 3 Evolutionary mechanisms: Mutation, selection, drift

From a population genetics perspective, we can define evolution as the change in the genetic
composition of a population over time. A number of mechanisms can effect this change,
including mutation, immigration or emigration, selection, drift. Mutations in a genome are
random occurrences, which provide a source of variation in a population. This variation can be
acted upon by natural selection, but the very fact that a new variant has arisen means that
evolution has occurred in the population, irrespective of any fitness differential conferred by
that variant. By our definition above, when a mutation occurs in a population, it changes the
allelic frequency or genetic composition of the population, thus resulting in evolution. Most
mutations are detrimental, while some are neutral, and a few may be advantageous.
Advantageous alleles that increase fitness are considered adaptations, and selection can act to
increase these alleles in the population. Even a neutral mutation, having no selective effects on
the population, may still rise in frequency by chance (genetic drift).

In lecture, several of the evolutionary forces that can cause new alleles to rise in frequency in a
population are discussed, including both natural selection and genetic drift. In this part of the
lab exercise, we will explore, comparing and contrasting, how these two forces in particular can
shape evolution.

Disentangling evolutionary forces in a hypothetical scenario: When is a mutation an

adaptation that leads to increased fitness and when is it simply a change?
To address this, we can use a case of contemporary evolution. Let’s consider some scenarios
related to an introduced spider species, Nephila clavata, to try to determine which evolutionary

8
force(s) is impacting the population of interest. Like many introduced invertebrate species,
these spiders likely arrived in the US as hitchhikers in cargo or shipping materials. They were
first observed in the US in 2013, in populations primarily discovered in Georgia. Since its
introduction in Georgia, has this population evolved?

Rather than focusing on spider silk production, as we did earlier in the lab, we are now looking
at similarities in COI (cytochrome oxidase I), a gene commonly used for species-level trees.
Researchers compared COI sequences from N. clavata samples from the Georgia populations
to COI samples from a wide geographical range of N. clavata populations to try to trace the
Georgia populations’ origin. See Fig. 4 Phylogeny of N. clavata individuals from a wide
geographic range, which depicts populations sharing brackets to be more similar to each other
than to populations outside those brackets.

Discuss: Which three possible source locations share the most similarity with the Georgia
populations?

Figure 4. Phylogeny of N. clavata individuals from wide geographic range. (Nephila

antipodiana serves as outgroup.)

Question 9. Which evolutionary forces are likely to be at play in this type of event? Why? What
effects do you think this would have on genetic variation in the introduced population?

Question 10. Why would information on geographical origin be useful for ecologists interested
in introduced species? (Provide two reasons.)

Simulating the effects of selection vs genetic drift

We have written an R-shiny app you can use to run simulations of evolution over time. Click this
link to go to the site and simulate the evolution of an allele (or use the following web address

9
https://eebprinceton.shinyapps.io/drift/). (Note that the app originally referenced a specific
viral variant, but the mechanics still apply to our spider scenarios.)

In the app, you will see options to input the initial allele frequency, and to enter various values
for population size, number of generations, number of replicate populations (aka how many
times you want to run the simulation), and – where relevant – the selection coefficient. Don’t
use commas when entering population size.

For this portion of the lab, we will consider a fictitious population of spiders as we model the
effects of drift and selection. Assumption: Spiders are haploid for the chromosome carrying
these alleles.

Scenario: A hypothetical mutation in the AgSp1 silk protein called MadeThis1Up was initially
present in a spider population at a frequency of 0.3, but has since reached a frequency of A1 =
0.80 in spiders globally.

Can drift alone explain this rise in MadeThis1Up? Let’s run some simulations to determine
whether drift alone is sufficient to explain similar changes in allele frequencies.

Question 11A
Table 3. Drift Alone
Run the simulator at least five times (but it’s so quick you can easily run it more than five times)
with different parameters for population size ranging from 1-10,000 (your choice, but be sure
to include both large and small populations!). Run it with the slider set to 20 replicates and 500
generations. Record the number of times the MadeThis1Up allele is fixed or lost in the
population in your simulations.

Simulation Population MadeThis1Up Selection N Fixed N Lost

size starting coefficient (Freq = 1) (Freq=0)
frequency
Example 10 0.3 0 3 17
1 0.3 0
2 0.3 0
3 0.3 0
4 0.3 0
5 0.3 0

Question 11B. What do you observe? Could drift alone explain a rapid, global rise in the
mutation?

10
Question 12A
Table 4. Selection Alone
Run the simulations at least five times again (still with 20 replicates each and 500 generations).
Use a population size set of 10,000. Pick several values for the selection coefficient and record
the number of times that MadeThis1Up is fixed or lost.

Simulation Population MadeThis1Up Selection N Fixed N Lost

size starting coefficient (Freq=1) (Freq=0)
frequency
Example 10,000 0.3 0.02 16 0
1 10,000 0.3
2 10,000 0.3
3 10,000 0.3
4 10,000 0.3
5 10,000 0.3

Question 12B. What do you observe? Could selection alone explain a rapid, global rise in the
mutation?

Question 13A
Table 5. Drift + Selection
Run the simulations again for five times (still with 20 replicates each and 500 generations). This
time, vary your choice of population size and the selection coefficient. Record the number of
times that MadeThis1Up is fixed or lost.

Simulation Population MadeThis1Up Selection N Fixed N Lost

size starting coefficient (Freq=1) (Freq=0)
frequency
Example 500 0.3 0.02 12 2
1 0.3
2 0.3
3 0.3
4 0.3
5 0.3

Question 13B. What would you infer about the relative roles of drift vs selection in the rise of
this allele? What aspects, if any, might these simulations be overlooking?

Question 14. Consider again the populations of Nephila clavata, introduced in 2013 in low
numbers to Georgia. Use what you’ve learned about drift and selection in these simulations to

11
discuss some ecological scenarios in which drift and selection might have influenced the
evolution of N. clavata.

References
Blackledge, Todd & Todd, & Scharff, Nikolaj & Coddington, Jonathan & Tamas, Szuts & John,
Wenzel & Hogenesch, John & Hayashi, & Cheryl, & Agnarsson, Ingi. (2009). Reconstructing Web
Evolution and Spider Diversification in the Molecular Era. Proceedings of the National Academy
of Sciences. 106: 5229-5234

Conserved Domain Database Spidroin_N [Internet]. Bethesda (MD): National Library of

Medicine (US), National Center for Biotechnology Information; [2021] - [cited 2022 Aug 01].
Available from: https://www.ncbi.nlm.nih.gov/Structure/cdd/

Fernandez et al. (2018). Phylogenomics, diversification dynamics, and comparative

transcriptomics across the Spider Tree of Life. Current Biology 28: 1489– 1497

Hoebeke ER, Huffmaster W, Freeman BJ. (2015) Nephila clavata L Koch, the Joro Spider of East
Asia, newly recorded from North America (Araneae: Nephilidae). PeerJ 3:e763

Scharff N et al. (2019) Phylogeny of the orb-weaving spider family Araneidae (Araneae:
Araneoidea). Cladistics 36: 1-21

Vollrath F and Selden P. (2007) The Role of Behavior in the Evolution of Spiders, Silks and Webs.
Annual Review of Ecology, Evolution, and Systematics 38: 819-46

Portions of this lab, including the allele frequency simulator, were modified from previous lab
exercises written by Sarah Kocher.

12