Phylogenetics Notes

Phylogenetics

• Phylogenetics is the study of the evolutionary relationships among organisms.
• Oliver A. Ryder (1986) highlighted the importance of preserving adaptive genetic variation, prompting consideration of which gene pools to prioritize.
• COSEWIC (2010) emphasizes the need to consider units below the species level (subspecies, varieties, or genetically distinct populations) in conservation efforts, as reflected in the Species at Risk Act.

Taxonomy & Systematics

• Taxonomy involves the identification and naming of taxa.
• Systematics focuses on assessing evolutionary relationships and phylogeny.
• Conservation decisions rely on the correct identification of species.

Identifying Species

• Failure to recognize species can lead to extinction, especially in the case of cryptic species.
• Over-splitting (recognizing more species than there actually are) can drain resources.
• Hybridization can result in outbreeding depression.

Approaches to Identifying Taxa

• Phenetics (Anagenesis): Groups organisms by phenotype, based on the similarity of current traits, resulting in a phenogram.
• Cladistics (Cladogenesis): Groups organisms by common ancestry, based on sharing derived traits, resulting in a cladogram or phylogeny.

Evolutionary Classification

• Modern taxonomy combines cladistics and phenetics to create an 'Evolutionary Classification'.
• Groups are initially classified by phylogeny (common ancestry).
• Phenotypically divergent groups are often recognized as separate taxa.
• Phylogeny describes the history of descent of taxa, showing the branching of species from common ancestors, often represented as a tree or dendrogram.
• The only figure in Darwin's "On the Origin of Species" is a tree depicting evolution.

Phylogenetic Analysis

• Describes the evolutionary relationships between species.
• Determines the order of speciation events.
• Estimates timing of speciation events.

Applications of Phylogenetic Analysis

• Epidemiology - Zoonotics: Understanding the transmission of diseases from animals to humans.
• Predicting Gene Function: Inferring the function of genes based on their evolutionary relationships.
• Evolutionary Biology: Studying the processes of evolution.
• Conservation:
  • Defining units of conservation.
  • Conservation triage - deciding which species to prioritize for conservation efforts.

Phylogenetic Network Analysis of SARS-CoV-2 Genomes

• A phylogenetic network analysis of 160 complete human SARS-CoV-2 genomes identified three central variants (A, B, and C).
• Variant A is considered the ancestral type, based on the bat outgroup coronavirus.
• Types A and C are found in significant proportions outside East Asia (Europeans and Americans).
• Type B is most common in East Asia.
• The ancestral genome of type B appears not to have spread outside East Asia without first mutating into derived B types, suggesting founder effects or resistance.
• The network analysis can trace routes of infections.
• Mutations:
  • Type B is derived from A by two mutations: synonymous mutation $T8782C$ and nonsynonymous mutation $C28144T$ (leucine to serine).

EDGE (Evolutionarily Distinct & Globally Endangered)

• Focuses on threatened species that represent a significant amount of unique evolutionary history.
• GE (Global Endangerment) score: based on IUCN (International Union for Conservation of Nature) category.
• ED (Evolutionary Distinctiveness) score: based on phylogeny.

Terminology

• Homologous genes: evolved from a common ancestor.
• Orthologous genes: evolved from a common ancestor by a speciation event; retain function.
• Paralogous genes: evolved from a common ancestor by duplication; different function.

Morphological vs. Molecular Phylogenies

• Mid-1950s: phylogenies constructed based on opinion.
• Classical phylogenetic analysis: used morphological features.
• Modern methods use DNA and amino acid sequences.
• Based on analysis of homologous sequences in different species.

Phylogenetic Tools

• DNA-DNA hybridization: Measures the heat required to split hybrid DNA (Sibley-Ahlquist used this for bird taxonomy).

Advantages of Phylogenetics

• DNA and AA sequences are strictly heritable units.
• Unambiguous description of molecular characters and states.
• Amenable to mathematical modelling.
• Homology assessment is easy.
• Distance relationships may be resolved.
• Huge amount of data is available.

mtDNA

• Represents a small fraction of the organismal genome size.
• Elevated mutation rate generates signal about population history over short time frames.
• Inheritance is clonal, simplifying within-species variation data.
• d-loop or control region and Cytochrome b are commonly used.
• Molecular clock considerations.

Gene Trees vs. Species Trees

• Genes can result in different phylogenies.

• Gene trees are often different from the true species phylogeny.

• Different gene phylogenies can arise due to:

  • lineage sorting
  • sampling error of individuals or populations
  • natural selection
  • introgression following hybridization

• Phylogeny of the species can be inferred from the gene tree.

• There is no unique universal phylogenetic tree.

Basic Steps in Phylogenetic Tree Construction

1. Identify & acquire homologous sequences.
2. Align the sequences.
3. Estimate the tree.
4. Draw the tree.
5. Assess the reliability of the tree.

Step 1: Identify & Acquire Homologous Sequences

• Primer extension reactions are used in sequencing.

Next Generation Sequencing (NGS)

• Databases for sequences and structures:
  • Nucleotide sequences: GenBank (NCBI), EMBL, DDBJ
  • Protein sequences: SWISS-PROT, PIR-International, OWL, NRDB, PROSITE, PRINTS, Pfam
  • Macromolecular structures: Protein Data Bank (PDB), Nucleic Acids Database (NDB), HIV Protease Database, ReLiBase, PDBsum, CATH, SCOP, FSSP
  • Genome sequences: Entrez genomes, GeneCensus, COGs
• Integrated databases: InterPro, Sequence retrieval system (SRS), Entrez

The International Nucleotide Sequence Database Collaborative (INSDC)

• A longstanding global alliance of biological data archives.
• Collects and disseminates DNA and RNA sequences.
• Three partners establish formats for data, metadata, and protocols:
  • DNA Data Bank of Japan (DDBJ)
  • European Nucleotide Archive (EMBL)
  • GenBank (NCBI)

Bioinformatics & Homology

• Find related sequences by searching against an online database (NCBI, EMBL, DDBJ).
• NCBI uses BLAST (Basic Local Alignment Search Tool).
• BLAST uses a sequence of interest as a “query” to search the database.
• Input sequence directly or enter accession number.
• E-value < 0.01 is considered a good match.

Step 2: Align the Sequences

• Mutation is the primary source of variation.

  • Substitutions: replacement of one nucleotide with another.
  • Recombinations: exchange of a sequence from one homologous chromosome to another.
  • Inversions: rotation by $180^{\circ}$ of a DNA segment.
  • Deletions: the loss of one or more nucleotides.
  • Insertions: the addition of one or more nucleotides.

• INDELS (insertions or deletions).

Sequence Alignment

• All sequences on a tree are homologous (true for species, but not always for genes).
• Phylogenetic methods assume that sequences in a column are homologous.
• Without INDELS, sequences can be written one above the other.
• INDELS change sequence length and shift positions of genes.
• Alignment introduces gaps to shift bases back to homologous positions.
• The tree can be no better than the alignment.

ClustalW Options

• Alignment with and without gaps

Step 3: Estimating the Tree

• A tree consists of tips or terminal nodes, branches, nodes, and a root.
• Sister taxa share a recent common ancestor.
• A polytomy occurs when data cannot resolve which taxa share a more recent common ancestor.
• Branches can be rotated at nodes without changing the relationships represented on the trees.
• (A, B, C) is a monophyletic group; it is sister to (D, E)
• (A, B, C, D) is paraphyletic → E is excluded

Rooted vs. Unrooted Trees

• Unrooted trees show evolutionary relationships but not the direction of evolution.
• A root is an interior node from which all sequences/taxa descend.
• An outgroup is used to determine the root; it is more distantly related to the ingroup sequences than the ingroup are to each other (often a sister taxa).

Methods for Estimating Trees

• Formula for calculating number of trees with taxa: $3 \text{ taxa} = 1 \text{ tree}$, $4 \text{ taxa} = 3 \text{ trees}$, $5 \text{ taxa} = 15 \text{ trees}$, $6 \text{ taxa} = 105 \text{ trees}$, $8 \text{ taxa} = 135,135 \text{ trees}$, $10 \text{ taxa} = 34,000,000 \text{ trees}$, $50 \text{ taxa} = 3 \times 10^{74} \text{ trees}$

Two Approaches to Reconstruction

1. Distance-based:
   • Input is a matrix of distances between species.
   • Neighbor-joining is the most common method and doesn’t assume a molecular clock (also UPGMA).
2. Character-based:
   • Examines all characters (AAs or DNA).
   • Parsimony chooses a tree that minimizes the number of changes required to explain the data.
   • Maximum likelihood, under a model of sequence evolution, finds the tree which gives the highest likelihood of the observed data.
   • Trees based upon both approaches are often reported.

Neighbor Joining Method

• When two sequences are similar, they likely originate from the same ancestor.
• Sequence similarity can approximate evolutionary distances.
• The more time has passed since a pair of sequences have diverged from a common ancestor, the more different they will be.
• One lineage may have evolved faster, or multiple substitutions may have occurred.

Advantages

• Fast and suited to large datasets and bootstrapping.
• Permits lineages with largely different branch lengths.
• Permits correction for multiple substitutions.

Disadvantages

• Sequence information is reduced.
• Gives only one possible tree.
• Strongly dependent on the model of evolution used (for divergent sequences).

Neighbor Joining - Algorithmic Method

• Uses a specific set of calculations to estimate a tree.
• Starting with a multiple alignment, calculate for each pair the distance or fraction of differences and write that to a distance matrix.

Distance Matrix

• Example:

  | | H | C | O |
  | :-------- | :-: | :-: | :-: |
  | Human (H) | - | 1 | 3 |
  | Chimp (C) | 1 | - | 2 |
  | Orang (O) | 3 | 2 | - |

Multiple Substitutions

• As two sequences diverge, each substitution will increase the differences between the two lineages.
• As differences accumulate, it’s more likely a substitution will occur at the same site.
• The number of observed differences almost always underestimates the actual amount of change.
• A variety of models are used to estimate corrected differences.

Models for Nucleotide Substitution

• Allows corrected distances, without underestimating.

  • Juke-Cantor (JC): all mutations are equally likely.
  • Kimura 2-parameter (K2P): transitions are more likely than transversions.
  • Felsenstein 84 (F84): K2P plus unequal base frequencies.
  • General Time Reversible (GTR): most general usable model.

Example of Neighbor Joining Calculation

Step 1

• Calculate the net divergence $r(i)$ for each OTU from all other OTUs (Total of 6 OTUs, $N=6$).
  • $r(A) = 5 + 4 + 7 + 6 + 8 = 30$

Step 2

• Adjusted distance (D) between each pair of nodes. Pairwise distance in starting matrix and divergence of each node.
  • For $D(A, B)$, $D(A, B) = d(A, B) - [r(A) + r(B)]/(N - 2) = 5 - (30 + 42)/(6 - 2) = -13$

Step 3

• Using the new matrix, find the closest pair of taxa $i, j$. Consider the lowest distance and assign $u$ as the connecting node for that pair.
• Branch length is calculated using the formula.
  • $S(i, u) = d(i, j)/2 + (r(i) - r(j))/2(n - 2)$
  • $S(j, u) = d(i, j) - S(i, u)$
• Closest as per matrix D is: AB = -13. The distance from U to A, and U to B is calculated as:
  • $S(A, U) = d(A, B)/2 + (r(A) - r(B))/2(n - 2) = 5/2 + (30 - 42)/2(6 - 2) = 1$
  • $S(B, U) = d(A, B) - S(A, U) = 5 - 1 = 4$
    Where, $d(A, B) = 5, r(A) = 30, r(B) = 42 \text{ and } n = 6$

Step 4

• Calculate the new distances from U to all other OTUs:
  • $d(U, C) = [d(A, C) + d(B, C) - d(A, B)]/2 = [4 + 7 - 5]/2 = 3$
• Repeat steps 1 to 4 using the new matrix of distances in every round

How Reliable Is My Tree?

• Non-parametric bootstrapping is commonly used to assess the uncertainty of tree reconstructions.
• Disturbs the observed data, composition of alignment columns Checks if the reconstructed trees are similar to the original.
• One bootstrap replicate randomly draws columns from the original alignment with replacement, repeating until the replicate contains as many columns as the original.

Bootstrapping

• $n$ is the number of bootstrap replicates obtained (typically 100 – 1,000).
• Produces $n$ bootstrap trees.
• Bootstrap values correspond to the relative frequency at which a node occurs in the bootstrap replicates.
• Typically ignore nodes with <60% support.