What you'll investigate
Paleontologists working in Liaoning Province, China, have found a new fossil. Rare soft tissue survived, and DNA from it was used to sequence three genes. Your job is to decide where this species belongs on a cladogram, first from its body structure and then from its DNA.
You will:
- Fill in a table of physical traits and use it to build a cladogram of six animal groups.
- Study the fossil and make a first hypothesis about where it fits.
- Search each fossil gene against a sequence database with an in-page version of BLAST. You will read scores, E-values, percent identity and the alignments themselves.
- Use the DNA evidence to keep or change where you placed the fossil.
- Design your own sequence comparison and test whether DNA differences build up like a "molecular clock."
- EVO-3.B Describe the types of evidence that can be used to infer an evolutionary relationship.
- EVO-3.C Explain how a phylogenetic tree and/or cladogram can be used to infer evolutionary relatedness.
- SYI-3.A Explain the connection between variation in the number and types of molecules within cells and an organism's ability to survive and/or reproduce in different environments.
- Science Practice 2.D Represent relationships within a biological model.
This is a computer investigation, so there are no lab hazards. The sequences in this page's database are simulated so that the whole lab works offline. You can try the real NCBI BLAST as an extension at the end.
Cladograms, shared genes, and BLAST
- A cladogram is a hypothesis about relatedness: groups that join at a more recent node share a more recent common ancestor.
- Cladograms are built from shared derived characters, traits that appeared in an ancestor and were passed to all its descendants.
- Most modern trees use molecular evidence: the longer two lineages have been apart, the more their DNA differs.
- BLAST finds local alignments and reports a bit score, an E-value (chance matches expected) and percent identity.
- A low E-value (below about 1 × 10−4) means a hit is probably a real homolog, not a coincidence.
A cladogram (a type of phylogenetic tree) shows a hypothesis about how species are related. The tips are species or groups. Each branch point (node) stands for a common ancestor. Two groups that join at a more recent node share a more recent common ancestor than groups that join closer to the root. You can rotate the branches around any node without changing what the tree says. Only the branching pattern matters.
Traditional cladograms are built from shared derived characters. These are traits that appeared in an ancestor and were passed to all of its descendants. On the cladogram below, each trait is marked where it first evolved. Every group above that mark has the trait. A trait shared by many groups usually evolved earlier than a trait found in only a few.
Molecular evidence and the molecular clock
Today, most cladograms are built mainly from molecular evidence. When a population splits into two species, each lineage gathers its own mutations. The longer two lineages have been apart, the more differences build up in their DNA. So the percentage of matching bases between two versions of a gene tells you roughly how recently the species shared an ancestor. For example, humans and chimpanzees share more than 95% of their DNA, and humans and fruit flies share about 60%.
BLAST: scores, E-values and percent identity
Comparing sequences by hand is slow. BLAST (Basic Local Alignment Search Tool) does it in seconds. You give it a query sequence. It finds short exact "seed" matches in every database sequence, then extends each seed into a full local alignment, allowing mismatches and gaps. Each alignment gets a score:
The raw score is converted to a bit score, which can be compared between searches. The E-value is the number of alignments this good that you would expect to find by chance in a database of that size:
The higher the score, the closer the alignment. The lower the E-value, the less likely the match is a coincidence. A hit with E < 1 × 10−4 is usually treated as a real, related sequence (a homolog). Percent identity is the number of matching positions divided by the length of the alignment. Query cover is the percent of your query that took part in the alignment.
Organism A ATGATCCAGT
| || ||||
Organism B ACGACTCAGT 7 of 10 match → 70% identity
DNA is only one line of evidence. Fossils, anatomy, embryology and protein sequences can all support or challenge a tree. A good cladogram is a hypothesis that fits the most evidence.
What this simulation models
- Process: Molecular evolution. DNA changes slowly over time, so species with a more recent common ancestor have more similar genes. A BLAST-style alignment finds and ranks the closest matches.
- Included: Three genes evolved along a tree of real divergence dates, each at its own speed (slow, medium, fast), with faster change at third codon positions, some sites that never change, and rare insertions or deletions. Hits are scored as bit scores and E-values.
- Simplified or left out: The sequences are simulated, not from GenBank, and the database holds about 20 species, not millions. Real fossil DNA this old would not survive. Real searches return messier, more surprising hits.
Before you start
Getting Started, Table 1 (plants: 1 = has the trait): Mosses 0 0 0 · Ferns vascular tissue only · Pine trees vascular tissue + seeds · Flowering plants vascular tissue + seeds + flowers.
| Species (vs. human) | GAPDH gene similarity (%) | GAPDH protein similarity (%) |
|---|---|---|
| Chimpanzee (Pan troglodytes) | 99.6% | 100% |
| Dog (Canis lupus familiaris) | 91.3% | 95.2% |
| Fruit fly (Drosophila melanogaster) | 72.4% | 76.7% |
| Roundworm (Caenorhabditis elegans) | 68.2% | 74.3% |
Build a cladogram from physical traits
Before you begin: Read each organism card, then mark only traits supported by the evidence on that card.
Need a quick science reminder?
A derived trait is a characteristic that evolved in a lineage. Mark traits actually shown in the evidence; don't infer missing traits from an organism's name.
Pick a data set. Read the organism cards, then tick every trait each organism has. Check your table, then group organisms that share a derived trait to build the tree. Start with the most widely shared traits.
1 · Trait table
2 · Group organisms into clades
Controls
Tip
Tip: a trait every organism shares goes at the root. Select all of them and choose that trait.
3 · Observe the fossil and make a hypothesis
Observations from the specimen:
- A backbone runs into a long, bony tail made of many vertebrae.
- The jaws hold small, sharp, curved teeth.
- The forelimbs end in three clawed fingers. The thumb is not opposable.
- Feather impressions line the arms, the legs and both sides of the tail.
- The hind legs are long, and it stood on two legs.
- A patch of soft tissue was preserved in the body cavity. This is where the DNA came from.
Search the fossil genes with mini-BLAST
| Term | Meaning |
|---|---|
| Query | Sequence you search with |
| Subject | Matched database sequence |
| Identity | Percent of aligned positions that match |
| Coverage | How much of the query aligns |
| E-value | Expected chance matches this strong in the searched database |
One strong match does not establish an evolutionary tree by itself.
Controls
More options
Full procedure
The real lab has you upload three saved gene files to NCBI BLAST. Here, the same three queries are built in. Run each gene with the settings below. For every search, read the hit list, open at least two alignments, and record the top hits. When you have run all three genes, look at the summary graph.
Best percent identity in each group. Each bar in the Graph tab shows the highest percent identity to the fossil query among that group's sequences, from your latest blastn search of each gene. A missing bar means no significant hit (E > 10).
Use the DNA evidence to place the fossil
Look again at your hypothesis from Part 1. Do the BLAST results support it? Decide where the fossil belongs and attach it to the cladogram you built. If your Part 1 tree is not finished, go back and finish the Figure 4 data set first.
BLAST your own question
Controls
More options
Full procedure
Scientists start with a question about a gene or a group of organisms. Choose any organism's version of gene A, B or C as your query, pick which organisms to search, and look for a pattern. Here is one testable idea. If mutations build up at a fairly steady rate, percent identity should drop steadily as the time since two species' common ancestor increases. This idea is called a molecular clock. Divergence times from the fossil record are given for each pair. You can also paste a sequence of your own (A, C, G and T only).
What does the evidence say?
How do I combine physical traits and DNA evidence?
- Compare the evidence types. Record which organisms share physical traits and which have more similar gene sequences.
- Read BLAST results carefully. Consider alignment length, percent identity, and statistical support; a high percent identity over a very short alignment may be less informative than a strong longer match.
- Support a placement. Explain where your fossil belongs on the cladogram using specific sequence and trait comparisons.
- Acknowledge uncertainty. One gene or one trait may not tell the whole evolutionary history. Describe any conflicts between the evidence types.
A closer sequence match supports a hypothesis of relatedness; it does not by itself prove direct ancestry.
Reflect and go further
Search NCBI's Gene database for a gene such as "human actin." Open its FASTA sequence and paste it into NCBI BLAST (nucleotide blast). Other genes to try: ATP synthase, catalase, GAPDH, keratin, myosin, Pax1, ubiquitin and zinc finger proteins. Compare the hit list and "Distance tree of results" with what you saw here.