7.91 / 7.36 / BE.490 Lecture #1 Feb. 24, 2004 Genome Sequencing & DNA Sequence Analysis Chris Burge What is a Genome? A genome is NOT a bag of proteins What’s in the Human Genome? Outline of Unit II: DNA/RNA Sequence Analysis Reading* 2/24 2/26 3/2 M Ch. 4 3/4 M Ch. 4 3/9 DNA Sequence Evolution 3/11 RNA Structure Prediction & Applications M Ch. 5 3/16 Literature Discussion TBA Genome Sequencing & DNA Sequence Analysis M Ch. 3 DNA Sequence Comparison & Alignment M Ch. 7 DNA Motif Modeling & Discovery Markov and Hidden Markov Models for DNA M Ch. 6 * M = Mount, “Bioinformatics: Sequence and Genome Analysis” Feedback to Instructor Examples from past years: ? Comic font looks stupid ? Burge uses too much genomics jargon ? Better synergy between Yaffe/Burge sections ? Asks questions to the class, student answers, but I didn’t hear/understand the answer… DNA vs Protein Sequence Analysis Protein Sequence Analysis DNA Sequence Analysis - emphasis on chemistry - emphasis on regulation - protein structure - RNA structure - selection is everywhere - signal vs noise (statistics) - multiple alignment - motif finding - comparative proteomics - comparative genomics - data: O(10^8) aa - data: O(10^10) nt Read your probability/statistics primer! Genome Sequencing & DNA Sequence Analysis ? The Language of Genomics ? Shotgun Sequencing ? DNA Sequence Alignment I ? Comparative Genomics Examples - Progress: genomes, transcriptomes, etc. - How to choose a mismatch penalty - PipMaker, Phylogenetic Shadowing Recent Media Attention Genomespeak Bork, Peer, and Richard Copley. " Genome Speak." Nature 409 (15 February 2001): 815. Learn to speak genomic In the following article, note the use of the following genomic terms: euchromatic, whole-genome shotgun sequencing, sequence reads, 5.11-fold coverage, plasmid clones, whole-genome assembly, regional chromosome assembly. Venter, JC, MD Adams, EW Myers, PW Li, RJ Mural, GG Sutton, HO Smith, … "The Sequence of The Human Genome." Science 291, no. 5507 (16 February 2001): 1304-51. Types of Nucleotides ? ribonucleotides ? deoxyribonucleotides ? dideoxyribonucleotides DNA Sequencing Adapted from Fig. 4.2 of “Genomes” by T. A. Brown, John Wiley & Sons, NY, 1999 Shotgun Sequencing a BAC or a Genome 200 kb (NIH) 3 Gb (Celera) Sequence, Assemble Sonicate, Subclone Subclones Shotgun Contigs What would cause problems with assembly? Shotgun Coverage (Poisson distribution) Sequence N reads, 500 bp each, from a 200kb BAC Coverage/read p = 500/200,000 = 0.0025 Total coverage C = Np Y = no. of reads covering the point x P(Y=k) = (N!/(N-k)!k!) p k (1-p) N-k ≈ e -c c k / k! . x P(Y=0)= e -c Examples: e -2 ≈ 0.14 e -4 ≈ 0.02 What could cause reality to differ from theory? Clickable Genomes Eukaryotes Protists Eubacteria S. cerevisiae Plasmodium E. coli S. pombe Giardia B. subtilis C. elegans … S. aureus Drosophila (several) … Anopheles (>100) Ciona Archaea Arabidopsis Methanococcus Human Sulfolobus Phages/Viruses Mouse … Lots Tetraodon (total of ~16) Fugu Organelles Zebrafish Lots Neurospora Aspergillus … Large-scale Transcript Sequencing Please see the following example article that uses large-scale transcript sequencing. Nature 420, no. 6915 (5 December 2002): 563-73. Okazaki, Y, M Furuno, T Kasukawa, J Adachi, H Bono, S Kondo, … "Analysis of The Mouse Transcriptome Based On Functional Annotation of 60,770 Full-length cDNAs." EST Sequencing dbEST release 022004 No. of public entries: 20,039,613 Summary by Organism - as of February 20, 2004 Homo sapiens (human) 5,472,005 Mus musculus + domesticus (mouse) 4,055,481 Rattus sp. (rat) 583,841 Triticum aestivum (wheat) 549,926 Ciona intestinalis 492,511 Gallus gallus (chicken) 460,385 Danio rerio (zebrafish) 450,652 Zea mays (maize) 391,417 Xenopus laevis (African clawed frog) 359,901 Hordeum vulgare + subsp. vulgare (barley) 352,924 Source: NCBI - http://ncbi.nlm.nih.gov * -omes and -omics Proteome Variome Transcriptome Genome Mass spec, Y2H, ? SNPs, haplotypes ESTs, cDNAs, microarrays Genome sequences Ribonome? Glycome ??? *Warning: some of the words on this slide may not be in Webster’s dictionary DNA Sequence Alignment I How does DNA alignment differ from protein alignment? Subject: Use BLASTN instead of BLASTP 1 ttgacctagatgagatgtcgttcacttttactgagctacagaaaa 45 |||| |||||||||||| | ||||||||||||||||||||||||| 403 ttgatctagatgagatgccattcacttttactgagctacagaaaa 447 Query: Nucleotide- nucleotide BLAST Web Server (BLASTN) DNA Sequence Alignment II Translating searches: translate in all possible reading frames search peptides against protein database (BLASTP) ttgacctagatgagatgtcgttcactttactgagctacagaaaa ttg|acc|tag|atg|aga|tgt|cgt|tca|ctt|tta|ctg|agc|tac|aga|aaa L T x M R C R S L L L S Y R K t|tga|cct|aga|tga|gat|gtc|gtt|cac|ttt|tac|tga|gct|aca|gaa|aa x P R x D V V H F Y x S T E tt|gac|cta|gat|gag|atg|tcg|ttc|act|ttt|act|gag|cta|cag|aaa|a D L D E M S F T F T E L Q K Also consider reading frames on complementary DNA strand DNA Sequence Alignment III Common flavors of BLAST: Program Query Database BLASTP aa aa BLASTN nt nt BLASTX nt (? aa) aa TBLASTN aa nt (? aa) TBLASTX nt (? aa) nt (? aa) PsiBLAST aa (aa msa) aa Which would be best for searching ESTs against a genome? DNA Sequence Alignment IV Which alignments are significant? Q: 1 S: Identify high scoring segments whose score S exceeds a cutoff x using dynamic programming. Scores follow an extreme value distribution: P(S > x) = 1 - exp[-Kmn e -λx ] For sequences of length m, n where K, λ depend on the score matrix and the composition of the sequences being compared (Same theory as for protein sequence alignments) ttgacctagatgagatgtcgttcacttttactgagctacagaaaa 45 |||| |||||||||||| | ||||||||||||||||||||||||| 403 ttgatctagatgagatgccattcacttttactgagctacagaaaa 447 Notes (cont) From M. Yaffe Lecture #2 ? The random sequence alignment scores would give rise to an “extreme value” distribution – like a skewed gaussian. ? Called Gumbel extreme value distribution For a normal distribution with a mean m and a variance σ, the height of the curve is described by Y=1/(σ√2π) exp[-(x-m) 2 /2σ 2 ] For an extreme value distribution, the height of the curve is described by Y=exp[-x-e -x ] …and P(S>x) = 1-exp[-e -λ(x-u)] where u=(ln Kmn)/λ Can show that mean extreme score is ~ log 2 (nm), and the probability of getting a score that exceeds some number of “standard deviations” x is: P(S>x)~ Kmne -λx. ***K and λ are tabulated for different matrices **** -λS For the less statistically inclined: E~ Kmne -2 -1 0.2 Yev 0.4 -4 4 0.4 B. Yn Probability values for the extreme value distribution (A) and the normal distribution (B). The area under each curve is 1. 0 1 2 X X A. 3 4 5 i DNA Sequence Alignment V How is λ related to the score matrix? λ is the unique positive solution to the equation*: ∑ p p j e λ s ij = 1 i i,j p = frequency of nt i, s ij = score for aligning an i,j pair What kind of an equation is this? (transcendental) What would happen to λ ? if we doubled all the scores? (reduced by half) What does this tell us about the nature of λ? (scaling factor) *Karlin & Altschul, 1990 DNA Sequence Alignment VI What scoring matrix to use for DNA? Usually use simple match-mismatch matrices: i j: A C G T A 1 m m m C m 1 m m s i,j : G T m m m m 1 m m 1 m = “mismatch penalty” (must be negative) DNA Sequence Alignment VII How to choose the mismatch penalty? Use theory of High Scoring Segment composition* High scoring alignments will have composition: q ij =p i p j e λs ij where q ij = frequency of i,j pairs (“target frequencies”) p , p = freq of i, j bases in sequences being compared i j What would happen to the target frequencies if we doubled all of the scores? *Karlin & Altschul, 1990