Multiple Sequence Alignment in BCB 444/544 Fall 07 - Lecture 11 - Prof. Drena Leigh Dobbs, Lab Reports of Bioinformatics

Download Multiple Sequence Alignment in BCB 444/544 Fall 07 - Lecture 11 - Prof. Drena Leigh Dobbs and more Lab Reports Bioinformatics in PDF only on Docsity! #11 - Multiple Sequence Alignment 9/14/07 BCB 444/544 Fall 07 Dobbs 1 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 1 BCB 444/544 Lecture 11 First BLAST vs FASTA Plus some Gene Jargon Multiple Sequence Alignment (MSA) #11_Sept14 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 2 √Mon Sept 10 - for Lecture 9/10 BLAST variations; BLAST vs FASTA, SW • Chp 4 - pp 51-62 √Wed Sept 12 - for Lecture 11 & Lab 4 Multiple Sequence Alignment (MSA) • Chp 5 - pp 63-74 Fri Sept 14 - for Lecture 12 Position Specific Scoring Matrices & Profiles • Chp 6 - pp 75-78 (but not HMMs) • Good Additional Resource re: Sequence Alignment? • Wikipedia: http://en.wikipedia.org/wiki/Sequence_alignment Required Reading (before lecture) 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 3 Assignments & Announcements - #1 Revised Grading Policy has been sent via email Please review! √Mon Sept 10 - Lab 3 Exercise due 5 PM: to: terrible@iastate.edu ?Thu Sept 13 - Graded Labs 2 & 3 will be returned at beginning of Lab 4 Fri Sept 14 - HW#2 due by 5 PM (106 MBB) Study Guide for Exam 1 will be posted by 5 PM 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 4 Review: Gene Jargon #1 (for HW2, 1c) Exons = "protein-encoding" (or "kept" parts) of eukaryotic genes vs Introns = "intervening sequences" = segments of eukaryotic genes that "interrupt" exons • Introns are transcribed into pre-RNA • but are later removed by RNA processing • & do not appear in mature mRNA • so are not translated into protein 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 5 Assignments & Announcements - #2 Mon Sept 17 - Answers to HW#2 will be posted by 5 PM Thu Sept 20 - Lab = Optional Review Session for Exam Fri Sept 21 - Exam 1 - Will cover: • Lectures 2-12 (thru Mon Sept 17) • Labs 1-4 • HW2 • All assigned reading: Chps 2-6 (but not HMMs) Eddy: What is Dynamic Programming 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 6 Chp 4- Database Similarity Searching SECTION II SEQUENCE ALIGNMENT Xiong: Chp 4 Database Similarity Searching • √Unique Requirements of Database Searching • √Heuristic Database Searching • √Basic Local Alignment Search Tool (BLAST) • FASTA • Comparison of FASTA and BLAST • Database Searching with Smith-Waterman Method #11 - Multiple Sequence Alignment 9/14/07 BCB 444/544 Fall 07 Dobbs 2 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 7 Why search a database? • Given a newly discovered gene, • Does it occur in other species? • Is its function known in another species? • Given a newly sequenced genome, which regions align with genomes of other organisms? • Identification of potential genes • Identification of other functional parts of chromosomes • Find members of a multigene family 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 8 FASTA and BLAST • FASTA • user defines value for k = word length • Slower, but more sensitive than BLAST at lower values of k, (preferred for searches involving a very short query sequence) • BLAST family • Family of different algorithms optimized for particular types of queries, such as searching for distantly related sequence matches • BLAST was developed to provide a faster alternative to FASTA without sacrificing much accuracy • Both FASTA, BLAST are based on heuristics • Tradeoff: Sensitivity vs Speed • DP is slower, but more sensitive 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 9 BLAST algorithms can generate both "global" and "local" alignments Global alignment Local alignment 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 10 BLAST - a Family of Programs: Different BLAST "flavors" • BLASTP - protein sequence query against protein DB • BLASTN - DNA/RNA seq query against DNA DB (GenBank) • BLASTX - 6-frame translated DNA seq query against protein DB • TBLASTN - protein query against 6-frame DNA translation • TBLASTX - 6-frame DNA query to 6-frame DNA translation • PSI-BLAST - protein "profile" query against protein DB • PHI-BLAST - protein pattern against protein DB • Newest: MEGA-BLAST - optimized for highly similar sequences http://www.ncbi.nlm.nih.gov/blast/producttable.shtml Which tool should you use? 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 11 Detailed Steps in BLAST algorithm 1. Remove low-complexity regions (LCRs) 2. Make a list (dictionary): all words of length 3aa or 11 nt 3. Augment list to include similar words 4. Store list in a search tree (data structure) 5. Scan database for occurrences of words in search tree 6. Connect nearby occurrences 7. Extend matches (words) in both directions 8. Prune list of matches using a score threshold 9. Evaluate significance of each remaining match 10. Perform Smith-Waterman to get alignment 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 12 1: Filter low-complexity regions (LCRs) ! ! ! " # $ $ $ % & = ' i i N n L L K ! ! log 1 Window length (usually 12) Alphabet size (4 or 20) Frequency of ith letter in the window • Low complexity regions, transmembrane regions and coiled-coil regions often display significant similarity without homology. • Low complexity sequences can yield false positives. • Screen them out of your query sequences! When appropriate! K = computational complexity; varies from 0 (very low complexity) to 1 (high complexity) e.g., for GGGG: L! = 4!=4x3x2x1= 24 nG=4 nT=nA=nC=0 P ni! = 4!x0!x0!x0! = 24 K=1/4 log4 (24/24) = 0 For CGTA: K=1/4 log4(24/1) = 0.57 This slide has been changed! #11 - Multiple Sequence Alignment 9/14/07 BCB 444/544 Fall 07 Dobbs 5 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 25 7: Extend matches in both directions DB Scan 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 26 7: Extend matches, calculating score at each step • Each match is extended to left & right until a negative BLOSUM62 score is encountered • Extension step typically accounts for > 90% of execution time L P P Q G L L Query sequence M P P E G L L Database sequence <word> 7 2 6 BLOSUM62 scores word score = 15 <--- ---> 2 7 7 2 6 4 4 HSP SCORE = 32 (High Scoring Pair) 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 27 8: Prune matches • Discard all matches that score below defined threshold 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 28 9: Evaluate significance • BLAST uses an analytical statistical significance calculation RECALL: 1. E-value: E = m x n x P m = total number of residues in database n = number of residues in query sequence P = probability that an HSP is result of random chance lower E-value, less likely to result from random chance, thus higher significance 2. Bit Score: S' = normalized score, to account for differences in size of database (m) & sequence length(n); Note (below) that bit score is linearly related to raw alignment score, so: higher S' means alignment has higher significance This slide has been changed! S'= (λ X S - ln K)/ln2 where: λ = Gumble distribution constant S = raw alignment score K = constant associated with scoring matrix For more details - see text & BLAST tutorial 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 29 10: Use Smith-Waterman algorithm (DP) to generate alignment • ONLY significant matches are re-analyzed using Smith-Waterman DP algorithm. • Alignments reported by BLAST are produced by dynamic programming 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 30 BLAST: What is a "Hit"? • A hit is a w-length word in database that aligns with a word from query sequence with score > T • BLAST looks for hits instead of exact matches • Allows word size to be kept larger for speed, without sacrificing sensitivity • Typically, w = 3-5 for amino acids, w = 11-12 for DNA • T is the most critical parameter: • ↑T ⇒ ↓ “background” hits (faster) • ↓T ⇒ ↑ ability to detect more distant relationships (at cost of increased noise) #11 - Multiple Sequence Alignment 9/14/07 BCB 444/544 Fall 07 Dobbs 6 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 31 Tips for BLAST Similarity Searches • If you don’t know, use default parameters first • Try several programs & several parameter settings • If possible, search on protein sequence level • Scoring matrices: PAM1 / BLOSUM80: if expect/want less divergent proteins PAM120 / BLOSUM62: "average" proteins PAM250 / BLOSUM45: if need to find more divergent proteins • Proteins: >25-30% identity (and >100aa) -> likely related 15-25% identity -> twilight zone <15% identity -> likely unrelated 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 32 Practical Issues Searching on DNA or protein level? In general, protein-encoding DNA should be translated! • DNA yields more random matches: • 25% for DNA vs. 5% for proteins • DNA databases are larger and grow faster • Selection (generally) acts on protein level • Synonymous mutations are usually neutral • DNA sequence similarity decays faster 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 33 BLAST vs FASTA • Seeding: • BLAST integrates scoring matrix into first phase • FASTA requires exact matches (uses hashing) • BLAST increases search speed by finding fewer, but better, words during initial screening phase • FASTA uses shorter word sizes - so can be more sensitive • Results: • BLAST can return multiple best scoring alignments • FASTA returns only one final alignment 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 34 BLAST & FASTA References • FASTA - developed first • Pearson & Lipman (1988) Improved Tools for Biological Sequence Comparison. PNAS 85:2444- 2448 • BLAST • Altschul, Gish, Miller, Myers, Lipman, J. Mol. Biol. 215 (1990) • Altschul, Madden, Schaffer, Zhang, Zhang, Miller, Lipman (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-402 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 35 BLAST Notes - & DP Alternatives • BLAST uses heuristics: it may miss some good matches • But, it’s fast: 50 - 100X faster than Smith-Waterman (SW) DP • Large impact: • NCBI’s BLAST server handles more than 100,000 queries/day • Most used bioinformatics program in the world!  But - Xiong says: "It has been estimated that for some families of protein sequences BLAST can miss 30% of truly significant matches." • Increased availability of parallel processing has made DP-based approaches feasible: • 2 DP-based web servers: both more sensitive than BLAST • Scan Protein Sequence: http://www.ebi.ac.uk/scanps/index.html Implements modified SW optimized for parallel processing • ParAlign www.paralign.org - parallel SW or heuristics 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 36 NCBI - BLAST Programs Glossary & Tutorials • http://www.ncbi.nlm.nih.gov/BLAST/ • http://www.ncbi.nlm.nih.gov/Education/BLASTinfo/glossary2.html • http://www.ncbi.nlm.nih.gov/Education/BLASTinfo/information3.html BLAST #11 - Multiple Sequence Alignment 9/14/07 BCB 444/544 Fall 07 Dobbs 7 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 37 Chp 5- Multiple Sequence Alignment SECTION II SEQUENCE ALIGNMENT Xiong: Chp 5 Multiple Sequence Alignment • Scoring Function • Exhaustive Algorithms • Heuristic Algorithms • Practical Issues 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 38 Multiple Sequence Alignments Credits for slides: Caragea & Brown, 2007; Fernandez-Baca, Heber &Hunter 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 39 Overview 1. What is a multiple sequence alignment (MSA)? 2. Where/why do we need MSA? 3. What is a good MSA? 4. Algorithms to compute a MSA 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 40 Multiple Sequence Alignment • Generalize pairwise alignment of sequences to include > 2 homologous sequences • Analyzing more than 2 sequences gives us much more information: • Which amino acids are required? Correlated? • Evolutionary/phylogenetic relationships • Similar to PSI-BLAST idea (not yet covered in lecture): use a set of homologous sequences to provide more "sensitivity" 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 41 What is a MSA? ATTTG- ATTTGC AT-TGC ATTTG ATTTGC ATT-GC ATTT-G- ATTT-GC AT-T-GC MSA Not a MSANot a MSA Why? 9/14/07BCB 444/544 F07 ISU Dobbs #11 - Multiple Sequence Alignment 42 Definition: MSA Given a set of sequences, a multiple sequence alignment is an assignment of gap characters, such that • resulting sequences have same length • no column contains only gaps ATTTG- ATTTGC AT-TGC ATTTG ATTTGC ATT-GC ATTT-G- ATTT-GC AT-T-GC YES NONO