genome assembly
play

Genome Assembly Sample Prepara1on Fragments Sequencing - PowerPoint PPT Presentation

Genome Assembly Sample Prepara1on Fragments Sequencing Reads ACGTAGAATACGTAGAA Assembly ACGTAGAATCGACCATG GGGACGTAGAATACGAC ACGTAGAATACGTAGAAACAGATTAGAGAG Con1gs Paired-End Reads Genomic


  1. Genome ¡Assembly ¡

  2. Sample ¡Prepara1on ¡ Fragments ¡ Sequencing ¡ Reads ¡ ACGTAGAATACGTAGAA Assembly ¡ ACGTAGAATCGACCATG GGGACGTAGAATACGAC ACGTAGAATACGTAGAAACAGATTAGAGAG… Con1gs ¡

  3. Paired-­‑End ¡Reads ¡ Genomic ¡segment ¡ Genome fragments Get ¡two ¡reads ¡from ¡ each ¡segment ¡ ~100 bp ~100 bp

  4. Read ¡Coverage ¡ C • Length ¡of ¡genomic ¡segment: ¡L ¡ • Number ¡of ¡reads: ¡n ¡ • Length ¡of ¡each ¡read: ¡l ¡ Coverage ¡ ¡C ¡= ¡n ¡l ¡/ ¡L ¡

  5. Fragment ¡Assembly ¡ • Cover ¡region ¡with ¡~7-­‑fold ¡redundancy ¡ • Overlap ¡reads ¡and ¡extend ¡to ¡reconstruct ¡the ¡ original ¡genomic ¡region ¡ ¡

  6. Challenges ¡in ¡Fragment ¡Assembly ¡ • Repeats: ¡a ¡ major ¡problem ¡for ¡fragment ¡assembly ¡ • > ¡50% ¡of ¡human ¡genome ¡are ¡repeats: ¡ ¡ ¡-­‑ ¡over ¡1 ¡million ¡ Alu ¡repeats ¡(about ¡300 ¡bp) ¡ ¡ ¡-­‑ ¡about ¡200,000 ¡LINE ¡repeats ¡(1000 ¡bp ¡and ¡longer) ¡ Repeat Repeat Repeat Green ¡and ¡blue ¡fragments ¡are ¡interchangeable ¡when ¡ ¡ assembling ¡repe11ve ¡DNA ¡

  7. Triazzle: ¡A ¡Fun ¡Example ¡ The ¡puzzle ¡looks ¡simple ¡ ¡ BUT ¡there ¡are ¡repeats!!! ¡ ¡ The ¡repeats ¡make ¡it ¡very ¡ difficult. ¡ ¡ Try ¡it ¡– ¡only ¡$7.99 ¡at ¡ www.triazzle.com ¡

  8. Repeat ¡Types ¡ • Low-­‑Complexity ¡DNA ¡(e.g. ¡ATATATATACATA…) ¡ • Microsatellite ¡repeats ¡ ¡ ¡ ¡ ¡(a 1 …a k ) N ¡where ¡k ¡~ ¡3-­‑6 ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡(e.g. ¡CAGCAGTAGCAGCACCAG) ¡ • Transposons/retrotransposons ¡ ¡ ¡ ¡ – SINE ¡ ¡ ¡ ¡Short ¡Interspersed ¡Nuclear ¡Elements ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡(e.g., ¡ Alu : ¡~300 ¡bp ¡long, ¡10 6 ¡copies) ¡ – LINE ¡ ¡ ¡ ¡Long ¡Interspersed ¡Nuclear ¡Elements ¡ ¡ ¡ ¡ ¡ ¡~500 ¡-­‑ ¡5,000 ¡bp ¡long, ¡200,000 ¡copies ¡ ¡ • Gene ¡Families ¡ ¡ ¡genes ¡duplicate ¡& ¡then ¡diverge ¡ • Segmental ¡duplicaCons ¡ ¡~very ¡long, ¡very ¡similar ¡copies ¡

  9. Fragment ¡Assembly ¡ • ComputaConal ¡Challenge: ¡ assemble ¡ individual ¡short ¡fragments ¡(reads) ¡into ¡a ¡single ¡ genomic ¡sequence ¡( “ superstring ” ) ¡ ¡ • Un1l ¡late ¡1990s ¡the ¡shotgun ¡fragment ¡ assembly ¡of ¡human ¡genome ¡was ¡viewed ¡as ¡ intractable ¡problem ¡ ¡ ¡ ¡

  10. Shortest ¡Superstring ¡Problem ¡ • Problem: ¡Given ¡a ¡set ¡of ¡strings, ¡find ¡a ¡shortest ¡ string ¡that ¡contains ¡all ¡of ¡them ¡ • Input: ¡ ¡Strings ¡ s 1 , ¡s 2 ,…., ¡s n ¡ • Output: ¡ ¡A ¡string ¡ s ¡that ¡contains ¡all ¡strings ¡ ¡ ¡ ¡ ¡s 1 , ¡s 2 ,…., ¡s n ¡as ¡substrings, ¡such ¡that ¡the ¡length ¡ of ¡ s ¡is ¡minimized ¡ ¡ • Complexity: ¡ ¡NP-­‑complete ¡ ¡ • Note: ¡ ¡this ¡formula1on ¡does ¡not ¡take ¡into ¡account ¡ sequencing ¡errors ¡

  11. Whole ¡Genome ¡Shotgun ¡Sequencing ¡ Genome ¡ Genome ¡amplified ¡and ¡sliced ¡into ¡ smaller ¡fragments ¡(>=600bp) ¡ Build ¡consensus ¡sequence ¡from ¡overlap ¡

  12. Overlap-­‑Layout-­‑Consensus ¡ ¡ Assemblers: ¡ARACHNE, ¡PHRAP, ¡CAP, ¡TIGR, ¡CELERA ¡ Overlap: ¡ ¡find ¡poten1ally ¡overlapping ¡reads ¡ Layout: ¡ ¡merge ¡reads ¡into ¡con1gs ¡and ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡con1gs ¡into ¡supercon1gs ¡ Consensus: ¡ ¡derive ¡the ¡DNA ¡sequence ¡ ..ACGATTACAATAGGTT.. and ¡correct ¡read ¡errors ¡

  13. Overlap ¡ • Each ¡read ¡is ¡compared ¡to ¡that ¡of ¡every ¡other ¡read, ¡in ¡both ¡the ¡ forward ¡and ¡reverse ¡complement ¡orienta1ons. ¡ ¡ • As ¡such, ¡the ¡overlap ¡computa1on ¡step ¡is ¡a ¡very ¡1me ¡intensive ¡ step ¡– ¡especially ¡if ¡the ¡set ¡of ¡reads ¡is ¡very ¡large. ¡ • For ¡example, ¡the ¡whole ¡genome ¡shotgun ¡assembly ¡of ¡ Drosophila ¡had ¡about ¡3 ¡x ¡10^6 ¡reads ¡of ¡500 ¡bases, ¡requiring ¡ roughly ¡10^13 ¡comparisons ¡(Deonier ¡ et ¡al., ¡ 2010). ¡ • Even ¡on ¡today's ¡computers, ¡running ¡that ¡many ¡comparisons ¡is ¡ imprac1cal, ¡so ¡seeded ¡algorithm ¡are ¡used ¡

  14. Overlapping ¡Reads ¡ • Sort ¡all ¡k-­‑mers ¡in ¡reads ¡ • Find ¡pairs ¡of ¡reads ¡sharing ¡a ¡k-­‑mer ¡ • Extend ¡to ¡full ¡alignment ¡– ¡throw ¡away ¡if ¡not ¡>95% ¡ similar ¡ TACA TAGATTACACAGATTAC T GA || ||||||||||||||||| | || TAGT TAGATTACACAGATTAC TAGA

  15. Finding ¡Overlapping ¡Reads ¡ Create ¡local ¡mul1ple ¡alignments ¡from ¡the ¡ overlapping ¡reads. ¡ TAGATTACACAGATTACTGA TAGATTACACAGATTACTGA TAG TTACACAGATTATTGA TAGATTACACAGATTACTGA TAGATTACACAGATTACTGA TAGATTACACAGATTACTGA TAG TTACACAGATTATTGA TAGATTACACAGATTACTGA

  16. Finding ¡Overlapping ¡Reads ¡ ¡ Correct ¡errors ¡using ¡mul1ple ¡alignment. ¡ • Find ¡loca1ons ¡where ¡there ¡is ¡a ¡devia1on ¡in ¡ which ¡1% ¡of ¡the ¡data ¡diverge ¡from ¡the ¡rest. ¡ • Make ¡those ¡posi1ons ¡agree ¡with ¡the ¡rest. ¡ TAGATTACACAGATTACTGA TAGATTACACAGATTACTGA TAG TTACACAGATTATTGA TAGATTACACAGATTACTGA TAGATTACACAGATTACTGA

  17. Build ¡the ¡Overlap ¡Graph ¡ • Overlap ¡graph: ¡the ¡nodes ¡represent ¡actual ¡ reads, ¡and ¡edges ¡represent ¡overlaps ¡between ¡ these ¡reads. ¡ ¡ • Thus, ¡the ¡genome ¡assembly ¡becomes ¡ equivalent ¡to ¡finding ¡a ¡path ¡through ¡the ¡graph ¡ that ¡visits ¡each ¡node ¡exactly ¡once ¡( i.e., ¡ a ¡ Hamiltonian ¡path ). ¡ 17 ¡

  18. An ¡overlap ¡graph. ¡Nodes ¡are ¡complete ¡reads ¡and ¡ edges ¡connect ¡reads ¡that ¡overlap. ¡Note ¡that ¡in ¡an ¡ actual ¡graph, ¡reads ¡and ¡overlaps ¡would ¡be ¡much ¡ larger. ¡ ¡ 18 ¡

  19. Layout ¡ • Finding ¡a ¡Hamiltonian ¡path ¡through ¡the ¡ overlap ¡graph ¡is ¡not ¡a ¡trivial ¡task. ¡ ¡ • In ¡order ¡to ¡decrease ¡the ¡size ¡of ¡the ¡graph, ¡the ¡ OLC ¡assembly ¡graph ¡is ¡simplified ¡in ¡the ¡layout ¡ stage, ¡where ¡segments ¡of ¡the ¡graph ¡are ¡ compressed ¡into ¡con1gs ¡ • Thus, ¡we ¡have ¡to ¡find ¡a ¡manner ¡to ¡decrease ¡ the ¡complexity ¡of ¡the ¡graph ¡

  20. Graph ¡Reduc1on ¡ • A ¡con1g ¡would ¡be ¡a ¡subgraph, ¡or ¡a ¡group ¡of ¡ nodes, ¡with ¡many ¡connec1ons ¡among ¡each ¡ other, ¡as ¡they ¡all ¡overlap ¡with ¡each ¡other ¡and ¡ refer ¡to ¡the ¡same ¡sequence ¡(A ¡and ¡B). ¡ ¡ • Once ¡a ¡subgraph ¡has ¡been ¡iden1fied, ¡these ¡ nodes ¡and ¡edges ¡are ¡compressed ¡into ¡one ¡ node, ¡or ¡a ¡con1g, ¡thereby ¡simplifying ¡the ¡ graph ¡(C) ¡ ¡ 20 ¡

  21. 21 ¡

  22. Separa1ng ¡Con1gs ¡ • There ¡are ¡two ¡classes ¡of ¡con1gs, ¡ unique ¡ conCgs ¡and ¡ repeat ¡conCgs . ¡ ¡ • Unique ¡con1gs ¡are ¡composed ¡of ¡reads ¡that ¡ can ¡be ¡unambiguously ¡assembled. ¡ ¡ • Repeat ¡con1gs ¡are ¡con1gs ¡with ¡an ¡abnormally ¡ high ¡read ¡coverage ¡or ¡connected ¡to ¡an ¡ abnormally ¡large ¡number ¡of ¡other ¡con1gs ¡ 22 ¡

  23. Separa1ng ¡Con1gs ¡ Normal density Too dense: Overcollapsed? Inconsistent links: Overcollapsed?

  24. Crea1ng ¡Scaffolds ¡ • Unique ¡con1gs ¡are ¡joined ¡into ¡larger ¡sequences, ¡ called ¡ scaffolds . ¡ ¡ • The ¡most ¡common ¡way ¡to ¡piece ¡con1gs ¡into ¡ scaffolds ¡is ¡through ¡ mate-­‑pair ¡informaCon . ¡ ¡ • With ¡mate-­‑pair ¡informa1on, ¡assemblers ¡can ¡ iden1fy ¡how ¡far ¡reads ¡and ¡unique ¡con1gs ¡should ¡ be ¡apart ¡from ¡each ¡other. ¡ ¡ – e.g. ¡if ¡a ¡2kb ¡fragment ¡of ¡a ¡genome ¡were ¡sequenced ¡ 100bp ¡on ¡each ¡end, ¡then ¡we ¡know ¡these ¡reads ¡and ¡ the ¡unique ¡con1gs ¡they ¡are ¡in ¡should ¡be ¡roughly ¡2kb ¡ apart. ¡ ¡ 24 ¡

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend