Intuition The big idea in one breathEvery organism carries a molecular diary — its DNA, RNA and proteins. Because these molecules are copied from parent to offspring with occasional mutations , two species that shared a recent common ancestor have more similar sequences than two that split long ago. So sequence differences act like a molecular clock : count the differences → estimate how closely related organisms are → build a family tree (phylogeny ) → classify.
Definition Core definitions
Phylogenetics ::= the study of evolutionary relationships between organisms (the branching "tree of life").Molecular phylogenetics ::= inferring those relationships by comparing sequences of molecules (DNA, RNA, amino acids) rather than physical (morphological) features.Phylogenetic tree ::= a branching diagram where tips = species, nodes = common ancestors, and branch length ∝ amount of evolutionary change.Molecular clock ::= the assumption that mutations accumulate at a roughly constant rate over time for a given molecule.
WHY molecules beat morphology (mostly):
Morphology can mislead — unrelated species evolve similar looks (convergence, e.g. dolphin vs shark).
Molecules give millions of comparable characters (every base pair), not a handful of bones.
Works even for organisms with no useful morphology to compare (bacteria, viruses).
Sequences are quantitative — you can count differences objectively.
Worked example The workflow, step by step
Choose a molecule. Pick a gene present in all the organisms you compare.
Why? You can only compare homologous (shared-ancestry) sequences.Classic choice: small subunit ribosomal RNA (16S in prokaryotes, 18S in eukaryotes) Carl Woese to discover the Archaea and propose the three-domain system.
Sequence & align. Line up the sequences so equivalent positions sit in the same column (insert gaps for insertions/deletions).
Why? Only aligned columns compare "the same" position across species.
Count differences between every pair of sequences → a distance matrix .
Why? Distance = a proxy for evolutionary time apart.
Build the tree so that more-similar sequences branch closer together.Root & interpret — read who shares a common ancestor with whom.
Intuition Why the correctione − 2 μ t e^{-2\mu t} e − 2 μ t D ≈ 2 μ t D\approx 2\mu t D ≈ 2 μ t twice (a mutation reversing an earlier one), so observed differences plateau below the true count — the curve saturates . That's why raw % difference underestimates deep divergences and why we use corrected distances.
Intuition Match the clock's speed to the question
Slow-evolving molecules (rRNA, essential enzymes like cytochrome c) → resolve ancient, deep splits (kingdoms, domains).Fast-evolving molecules (mitochondrial DNA, viral genes) → resolve recent, shallow splits (populations, closely related species).Why? A clock too slow shows no differences between close relatives; a clock too fast is fully saturated (all random) for distant ones.
Worked example Example 1 — reading a distance matrixSequence differences (number of differing bases):
A
B
C
A
–
2
8
B
2
–
9
C
8
9
–
Q: Which two are most closely related?
A: A & B (only 2 differences). C is the outgroup — it split off first.
Why this step? Smallest distance = fewest accumulated mutations = most recent common ancestor.
Worked example Example 2 — estimating divergence timeTwo species differ at D = 0.10 D = 0.10 D = 0.10 μ = 1 × 10 − 9 \mu = 1\times10^{-9} μ = 1 × 1 0 − 9 per year per lineage. Estimate the split time.
Using t ≈ D / ( 2 μ ) t \approx D/(2\mu) t ≈ D / ( 2 μ ) t ≈ 0.10 2 × 10 − 9 = 5 × 10 7 years = 50 million years t \approx \frac{0.10}{2\times10^{-9}} = 5\times10^{7}\ \text{years} = 50\ \text{million years} t ≈ 2 × 1 0 − 9 0.10 = 5 × 1 0 7 years = 50 million years Why this step? Both lineages have been changing since the split, so we divide by 2 μ 2\mu 2 μ (Corrected version: t = − ln ( 1 − D ) / ( 2 μ ) = − ln ( 0.9 ) / 2 × 10 − 9 ≈ 5.27 × 10 7 t=-\ln(1-D)/(2\mu)= -\ln(0.9)/2\times10^{-9}\approx 5.27\times10^7 t = − ln ( 1 − D ) / ( 2 μ ) = − ln ( 0.9 ) /2 × 1 0 − 9 ≈ 5.27 × 1 0 7
Worked example Example 3 — Woese's discoveryWhen 16S/18S rRNA was compared, some "bacteria" (methanogens, extreme halophiles) were more different from ordinary bacteria than expected. Their sequences formed a separate branch .
Result: life reclassified from 5 kingdoms into three domains — Bacteria, Archaea, Eukarya .
Why it mattered: molecules revealed a division invisible to the microscope.
Common mistake "More similar sequence always means more recently diverged."Why it feels right: intuitively, similar = closely related.
The flaw: if the two molecules evolve at different rates , or if saturation has erased changes, similarity can mislead. Also, plasmids/genes can be horizontally transferred (esp. in bacteria), so one gene's tree ≠ the organism's tree.
Fix: use conserved, vertically-inherited genes (like rRNA), apply distance corrections , and compare multiple genes .
Common mistake "Branch length in a tree = time."Why it feels right: longer branch → "further apart."
The flaw: branch length usually measures amount of change , not time — unless a strict clock is assumed. A fast-evolving lineage has a long branch in little time.
Fix: state whether the tree is a phylogram (change) or a chronogram (time).
Common mistake "You can compare any two sequences."Why it feels right: they're all DNA.
The flaw: only homologous, properly aligned positions are comparable. Comparing non-homologous genes gives noise.
Fix: verify homology and align first.
Recall Feynman: explain to a 12-year-oldImagine every animal keeps a secret notebook (its DNA) passed down from its parents. Each time it's copied, a tiny typo sneaks in. Cousins who split apart a long time ago have lots of typos that are different; cousins who split recently have almost the same notebook. So if we line up the notebooks and count the different letters, we can figure out who's closely related to whom and draw the whole family tree of life — even for tiny germs we can't tell apart just by looking!
Mnemonic Remember the pipeline"Some Aligned DNA Builds Trees" → S equence choose → A lign → D ifferences (distance) → B uild → T ree.
And for the clock: "Change Counts Clock Time" — more Changes ⇒ more Time (when rate is constant).
What is molecular phylogenetics? Inferring evolutionary relationships by comparing molecular sequences (DNA/RNA/protein) rather than morphology.
Why can molecules classify organisms better than morphology? They give millions of objective, countable characters and avoid being fooled by convergent evolution.
Which molecule did Carl Woese use and why? Small-subunit ribosomal RNA (16S/18S) — universal, essential, slow-evolving, so it works across all life.
What three domains did molecular data reveal? Bacteria, Archaea, Eukarya.
What is a molecular clock? The assumption that mutations accumulate at a roughly constant rate over time for a given molecule.
Derive the fraction of unchanged sites after time t. dP/dt = -μP ⇒ P = e^(−μt); fraction changed D = 1 − e^(−μt).
For two diverging species, why does the exponent use 2t? Both lineages accumulate change independently since the split, so total change path = 2t.
Approximate divergence time from distance D (small D). t ≈ D/(2μ).
Why does raw % difference underestimate deep divergences? Saturation — the same site mutates more than once, hiding earlier changes.
What must be true of sequences before you can compare them? They must be homologous (shared ancestry) and properly aligned.
Difference between a phylogram and a chronogram? Phylogram branch length = amount of change; chronogram branch length = time.
Fast vs slow evolving molecules — which resolves deep vs shallow splits? Slow (rRNA) → deep/ancient splits; fast (mtDNA) → recent/shallow splits.
Why can one gene's tree differ from the organism's true tree? Horizontal gene transfer and differing evolutionary rates between genes.
In a distance matrix, which pair is most closely related? The pair with the smallest number of differences.
Homologous gene e.g. 16S/18S rRNA
Archaea & 3-domain system
Intuition Hinglish mein samjhoDekho, molecular phylogenetics ka core idea simple hai: har organism ke andar ek "DNA diary" hoti hai jo parents se bachchon me copy hoti rehti hai, aur har copy me thoda-bahut typo (mutation) aa jata hai. Jo species haal hi me alag hue hain, unki sequences almost same hoti hain; jo bahut pehle alag hue, unme differences zyada. To bas sequences ko compare karke, differences count karke, hum bata sakte hain kaun kiska kitna close relative hai — aur isse family tree (phylogeny) banate hain, jo classification me use hota hai.
Molecular clock ka maths bhi easy logic pe khada hai: mutation constant rate μ \mu μ e − μ t e^{-\mu t} e − μ t D = 1 − e − 2 μ t D = 1 - e^{-2\mu t} D = 1 − e − 2 μ t D D D D ≈ 2 μ t D \approx 2\mu t D ≈ 2 μ t jitne zyada changes, utna zyada time since split. Ek dhyan dene wali baat: bahut purane relationships me "saturation" ho jata hai — ek hi site do baar mutate ho jati hai, to actual changes chhup jaate hain, isliye raw % difference thoda kam dikhata hai.
Ye topic isliye important hai kyunki morphology (bahar ki shakal) kabhi-kabhi dhoka de deti hai — dolphin aur shark dekhne me similar par relate nahi. Molecules objective hote hain, millions of characters dete hain, aur bacteria/virus jaise cheezon ke liye bhi kaam karte hain jinme dekh ke kuch pata nahi chalta. Isi technique se Carl Woese ne 16S rRNA compare karke Archaea discover kiya aur life ko teen domains — Bacteria, Archaea, Eukarya me divide kiya.
Exam tip: pipeline yaad rakho — Sequence choose karo, Align karo, Differences nikaalo, tree Build karo. Aur formula t ≈ D / ( 2 μ ) t \approx D/(2\mu) t ≈ D / ( 2 μ )
