3 Gene Finding via HMMsGene finding can be interpreted as a d.p. approach that threads genomic sequence through the states of a ‘gene’ HMM.Einit, Efin, Emid,I, IG (intergenic)IGIEfinEmidNote: all links are notshown hereEiniti
4 Generalized HMMs, and other refinements A probabilistic model for each of the states (ex: Exon, Splice site) needs to be describedIn standard HMMs, there is an exponential distribution on the duration of time spent in a state.This is violated by many states of the gene structure HMM. Solution is to model these using generalized HMMs.
6 Generalized HMM for gene finding Each state also emits a ‘duration’ for which it will cycle in the same state. The time is generated according to a random process that depends on the state.
7 Forward algorithm for gene finding qkjiDuration Prob.: Probability that you stayedin state qk for j-i+1 stepsEmission Prob.: Probability that you emitted Xi..Xj instate qk (given by the 5th order markov model)Forward Prob: Probability that you emitted i symbols and endedup in state qk
8 De novo Gene prediction: Summary Various signals distinguish coding regions from non-codingHMMs are a reasonable model for Gene structures, and provide a uniform method for combining various signals.Further improvement may come from improved signal detection
9 DNA Signals 5’ UTR intron exon 3’ UTR Acceptor Donor splice site Coding versus non-codingSplice SignalsTranslation startATG5’ UTRintronexon3’ UTRAcceptorDonor splice siteTranscription startTranslation start
10 DNA signal example:The donor site marks the junction where an exon ends, and an intron begins.For gene finding, we are interested in computing a probabilityD[i] = Prob[Donor site at position i]Approach: Collect a large number of donor sites, align, and look for a signal.
11 PWMs 321123456 AAGGTGAGT CCGGTAAGT GAGGTGAGG TAGGTAAGG Fixed length for the splice signal.Each position is generated independently according to a distributionFigure shows data from > 1200 donor sites
12 Improvements to signal detection Pr[GGTA] is a donor site?0.5*0.5Pr[CGTA] is a donor site?Is something wrong with this explanation?GGTACGTG
13 MDD PWMs do not capture correlations between positions Many position pairs in the Donor signal are correlated
14 Maximal Dependence Decomposition Choose the position i which has the highest correlation score.Split sequences into two: those which have the consensus at position i, and the remaining.Recurse until <Terminating conditions>Stop if #sequences is ‘small enough’
16 Gene prediction: Summary Various signals distinguish coding regions from non-codingHMMs are a reasonable model for Gene structures, and provide a uniform method for combining various signals.Further improvement may come from improved signal detection
19 Comparative methodsGene prediction is harder with alternative splicing.One approach might be to use comparative methods to detect genesGiven a similar mRNA/protein (from another species, perhaps?), can you find the best parse of a genomic sequence that matches that target sequenceYes, with a variant on alignment algorithms that penalize separately for introns, versus other gaps.There is a genome sequencing project for a different Hirudo species. You could compare the Hirudo ESTs against the genome to do gene finding.
20 Comparative gene finding tools Procrustes/Sim4: mRNA vs. genomicGenewise: proteins versus genomicCEM: genomic versus genomicTwinscan: Combines comparative and de novo approach.Mass Spec related?Later in the class we will consider mass spectrometry data.Can we use this data to identify genes in eukaryotic genomes? (Research project)
21 DatabasesRefSeq and other databases maintain sequences of full-length transcripts/genes.We can query using sequence.
22 Course Sequence Comparison (BLAST & other tools) Protein Motifs: Gene findingSequence Comparison (BLAST & other tools)Protein Motifs:Profiles/Regular Expression/HMMsDiscovering protein coding genesGene finding HMMsDNA signals (splice signals)How is the genomic sequence itself obtained?ESTsProtein sequence analysis
23 Silly QuizWho are these people, and what is the occasion?
25 DNA Sequencing DNA is double-stranded The strands are separated, and a polymerase is used to copy the second strand.Special bases terminate this process early.
26 Sequencing A break at T is shown here. Measuring the lengths using electrophoresis allows us to get the position of each TThe same can be done with every nucleotide. Fluorescent labeling can help separate different nucleotides
27 Automated detectors ‘read’ the terminating bases. The signal decays after 1000 bases.
28 Sequencing Genomes: Clone by Clone Clones are constructed to span the entire length of the genome.These clones are ordered and oriented correctly (Mapping)Each clone is sequenced individually
29 Shotgun Sequencing Shotgun sequencing of clones was considered viable However, researchers in 1999 proposed shotgunning the entire genome.
30 LibraryCreate vectors of the sequence and introduce them into bacteria. As bacteria multiply you will have many copies of the same clone.
32 QuestionsAlgorithmic: How do you put the genome back together from the pieces? Will be discussed in the next lecture.Statistical?EX: Let G be the length of the genome, and L be the length of a fragment. How many fragments do you need to sequence?The answer to the statistical questions had already been given in the context of mapping, by Lander and Waterman.
34 LW statistics: questions As the coverage c increases, more and more areas of the genome are likely to be covered. Ideally, you want to see 1 island.Q1: What is the expected number of islands?Ans: N exp(-c)The number increases at first, and gradually decreases.
35 Analysis: Expected Number Islands Computing Expected # islands.Let Xi=1 if an island ends at position i, Xi=0 otherwise.Number of islands = ∑i XiExpected # islands = E(∑i Xi) = ∑i E(Xi)
36 Prob. of an island ending at i E(Xi) = Prob (Island ends at pos. i)=Prob(clone began at position i-L+1AND no clone began in the next L-T positions)
37 LW statistics Pr[Island contains exactly j clones]? Consider an island that has already begun. With probability e-c, it will never be continued. ThereforePr[Island contains exactly j clones]=Expected # j-clone islands