When all sequences belong to one clusterLecture 18Overview of PhylogenyLecturer: Prof. Serafim Batzoglou Scribe: Gaurav Garg([email protected]) ([email protected])1. What are Phylogenies?2. Inferring Phylogenies3. Background on Trees4. Phylogeny and Sequence Comparison5. Modeling EvolutionCS 262: Computational Genomics Spring 2002-2003Lecture 18Overview of PhylogenyLecturer: Prof. Serafim Batzoglou Scribe: Gaurav Garg ([email protected]) ([email protected])We have been talking about phylogeny throughout the course but never really looked at itin depth, so in this last lecture we intend to tie everything together by covering in a littlemore depth phylogeny and evolution. We believe that all life forms are connected throughan evolutionary tree. Also, evolution is something we want to understand on its ownregard, so in this age of high throughput genomics we hope genomics could help usunravel the various mysteries of life through sequence comparison. The figure on lastpage shows a small part of the evolutionary tree, of mammals, which are a tiny fraction ofall the organisms on the earth. For the mammals we have a pretty good idea of what theevolutionary tree should look like e.g. we know that we are closer to monkeys than torodents, similarly cats and dogs are closer to one another than to cows or pigs. We alsobelieve that we are closer to the rodents in terms of evolutionary time (not sequence) thanto dogs, cats or cows. But this is still a somewhat unclear question, we know pretty wellbut not quite 100%. In fact when we look at organisms other than mammals, like viruses,bacteria we do not have a good idea of their evolutionary relationships at all. So thismotivates us for the rest of the lecture.1. What are Phylogenies?Phylogenies are trees that show the history of life. A phylogeny tree shows the connectionbetween various organisms and weight of the branches in the tree give an idea of timebetween evolutions of different organisms. A node with two edges out of it is thecommon ancestor through which speciation occurs, e.g. humans have a common ancestorwith mice 80 million years ago. Beyond actual organisms particular sequence elementscan also have their own evolution history within a genome. We have already seen selfishDNA. Selfish DNA is repeating transposable element, i.e. a selfish DNA makes a copy ofit and inserts itself at a different position in the Genome. Now these elements haveessentially fooled the genome into believing that it actually exists at two places within it.This can arise just like life, by chance. So if we have a DNA that transposes itself thenmore copies of its will transpose themselves more and in this way we get evolution ofrepeats. Hence within a Genome we can get a phylogeny of repeats. We can similarlyhave phylogeny of genes that copy themselves by duplication. Across organisms aparticular gene family may have a whole history of how it evolved in the differentbranches of the tree. Hence we can conclude that phylogeny tries to answer the followingquestions. How do you establish relationships between (a) different organisms (b)between one family of elements within an organism or across organisms?Before we move further lets define the Orthologs and Paralogs. Orthologs are twoelements that have diverged because of speciation whereas Paralogs are two elements thathave diverged because of duplication. So if we have a gene and this gene get duplicatedand make two copies of it. And say we have two organisms; mouse and rat that evolvedafter the duplication then each one will have two copies of the gene. The two originals(between mouse and rat) are Orthologs, the two copies (between mouse and rat) are alsoOrthologs but an original and copy pair (between mouse and rat) is Paralogs. So bothoriginal versions of the gene in rat and mouse have come from the same gene, and bothcopies have come from the same copy version. But copy version in mouse and original inrat have not come from the same sequence, but they are very similar. They have actuallycome from a duplication of the original sequence before. Both orthology and paralogy aretogether called homology. Any two similar sequences are thus homologos, unless they aresimilar by some stroke of luck, which is very highly unlikely.2. Inferring PhylogeniesPhylogeny trees can be inferred in many different ways e.g. by looking at the morphologyof the organisms. E.g. According to morphology organisms with 4 legs could be groupedtogether but this is not true actually, the crocodile has 4 legs, dog has 4 legs and bat has 2legs, but dogs are closer to bats than to crocodiles because they are mammals. Somorphology cannot in general tell you a lot about the phylogeny of animals as comparedto molecular methods that do a sequence comparison. Take the following example ofspecies for which we do not know their phylogeny Orc: ACAGTGACGCCCCAAACGT Elf: ACAGTGACGCTACAAACGT Dwarf: CCTGTGACGTAACAAACGA Hobbit: CCTGTGACGTAGCAAACGA Human: CCTGTGACGTAGCAAACGAThe different colors show the alignment. It is more common for dwarfs, humans andhobbits to have a similar nucleotide than for humans and orcs or humans and elves. Butunlike this example it is not always easy to make this tree by the eye, so before weproceed further lets make a little background of the trees.3. Background on TreesEach edge of the tree can have length that specifies the time of evolution, or better theeffective time of evolution because some species evolve much faster than the otherspecies, e.g. rodents are believed to evolve much faster than other organisms. The treecan be rooted or unrooted in general. If the tree is unrooted than we don’t really knowwhere the common ancestor should be placed e.g. as shown in the figure we could placethe common ancestor (the root) at number of places. But there is no good way to figureout where to place this root.A tree with N leafs has 2N – 2 nodes if it is unrooted and has 2N – 1 nodes if it rooted.This is because our trees are all binary. We assume all our trees to be binary because it isvery highly unlikely that at exactly the same point three different species evolved.
View Full Document
Unlocking...