sequencing gap - Rice Genome Annotation Project

Library Construction• Break the genomic DNA from theorganism into many small fragments• The fragments are cloned into plasmidvectors• This collection of fragments is called alibraryTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Library Requirements1. Free vector should be at low or undetectable level.2. No chimeric clones. Chimeras occur two or more randomfragments from separate parts of the genome recombineand end up next to each other.3. The majority of the inserts should be of relatively uniformsize.4. Libraries need to be random and cover the wholegenome.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

DNA Shearing with HydroShearHydroShear Assembly Cleaning QC1 2 3 4Nebulized DNA vs HydroShear sheared DNASize Selection GelsTIGR7 ngQuant Gels Prior to Vector Ligation18ng 20ng 25 ngTHE INSTITUTE FOR GENOMIC RESEARCHPCR Products After 4 Wash Cycles1 BAC DNA, 2 sheared BAC DNA,3 sheared water, 4 sterile waterBACs assembled and checkedfor contaminant sequencesXSCQB – 3 seqs(2- Lotus japonicus, 1- Homo sapiens)XSCQC – 0XSCQD – 0XSCQE – 0

BstXI adaptor cloning systemDNA insertLigateBstXI adaptorCTTTCCAGCACAGAAAGGTCLigateComplementary to BstXI adaptorCTGGAAAGGTGTGACCTTTCVectorTIGRVectorplus insertTHE INSTITUTE FOR GENOMIC RESEARCH

Design Features of the BstXI AdaptorCloning SystemForward sequencing primerForward PCR primerrrnBT1BstXI site BstXI site Reverse sequencing primerReverse PCR primerrrnBT2ter2pHOS vector plus insertProriAmpRter1TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

pHOS Vector Features• The sequencing primer sites immediately flank the cloningsite to avoid excessive re-sequencing of vector DNA.• PCR primer sites are located immediately outside of theSequencing primer sites to allow PCR amplification fortemplate preparation.• The vector continues to have the regular, good vectorfeatures such as a gene for replication (origin of replication),antibiotic resistance and multiple cloning sites where wecan cut the circular plasmid and insert desired features.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Template Production• Transform clones intoE. coli• Plate and pick coloniesto isolate clones• Replicate and isolateplasmid for sequencing• 384 well HT plateprocessingTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Alkaline Lysis 384-well DNA Isolation MethodPick colonies into media, growFreezer copyCentrifuge to pellet cellsCell LysisDNA PurificationRe-suspend in TE/BlueAverage concentration30 - 40 ng/µLTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Template Production LaboratoryCurrent Capacity: 22,000,000 plasmids/yearExpansion Capacity: 54,000,000 plasmids/yearTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence Production• Sequencing Reactionsand Precipitations• Samples are loaded intothe capillaries on asequencing machine• Data Collection• Data AnalysisTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence Production LaboratoryCurrent Capacity: 40,000,000 sequences/yearExpansion Capacity: 100,000,000 sequences/yearTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Capillary array viewTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sample ElectropherogramTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Basecalling & Quality AssignmentsTIGRTHE INSTITUTE FOR GENOMIC RESEARCHPhred & TraceTuner•Read DNA sequencer traces•Call bases•Assign base quality values•Write basecalls and qualityvalues to output files.Warner Brothers, Inc.Lowering the error rate, averaging 40% - 50% fewer errors thanABI software independent of position in read, machine runningconditions, or sequencing chemistry

What are phred quality values?The quality value q assigned to a base call isdefined as:q = - 10 x log 10 (p)where p is the estimated error probabilityfor that base-call.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

ORA base-call having a probability of1/1000 of being incorrect is assigned aquality value of 30.ProbabilityQuality Value1/100 201/10 10TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

QC Tools1. Web based document control system including SOP’s andProcess Control Forms2. Receipt testing and lot control of sequencing reagents3. QC audits of reagents and protocols used byproduction teams4. Record of operators performing each step of the process5. Equipment calibration and preventive maintenance program6. QC/R&D group and QC representative on each teamTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Assembling the fragmentsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Celera Assembler• creates high confidence “uniquelyassembleable contigs” = unitigs• marks those that appear repetitive w.r.t.arrival-rate statistics (surrogates anddegenerates)• uses insert (clone-mate) information tobuild contigs & mark ambiguous unitigs(surrogates)TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Merging two sequencesoverlap (19 bases)overhang (6 bases)…AGCCTAGACCTACAGGATGCGCGGACACGTAGCCAGGACCAGTACTTGGATGCGCTGACACGTAGCTTATCCGGT…TIGRoverhangoverlap - region of similarity between regionsoverhang - un-aligned ends of the sequencesThe assembler screens merges based on:• length of overlap• % identity in overlap region• maximum overhang size.THE INSTITUTE FOR GENOMIC RESEARCH% identity = 18/19 % = 94.7%

Arrival rate statistics8x coverage, 800bp reads - a read every 100bp0 100 200 300 400 800A-stat = log (p(single copy) / p(two-copy))p(single copy) = ((pF/G)^k/k!) exp(-pF/G)p(two-copy) = ((2pF/G)^k/k!) exp(-2pF/G)TIGRTHE INSTITUTE FOR GENOMIC RESEARCHF - # fragmentsG - genome sizep/k - average arrival rate

Forward-reverse constraints• The sequenced ends are facing towards each other• The distance between the two fragments is known(within certain experimental error)clone lengthFRsequenced endsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Surrogate UnitigsA UB U CUAUBTIGRTHE INSTITUTE FOR GENOMIC RESEARCHC

Degenerate contigs• Too deep• Too few readsTIGR(NOT IN SCAFFOLDS)THE INSTITUTE FOR GENOMIC RESEARCH

GreedyTIGR Assembler• Build a rough map of fragmentoverlaps• Pick the largest scoring overlap• Merge the two fragments• Repeat until no more merges can bedoneTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Scaffolding• Given a set of non-overlapping contigsorder and orient them along a chromosomeII I III IVIIIIIIVTIGRTHE INSTITUTE FOR GENOMIC RESEARCHI

Clone-matesInsertFRIIIRFVectorIIIRFIIITIGRTHE INSTITUTE FOR GENOMIC RESEARCHFR

Linking information• Overlaps• Mate-pair links• Similarity links• Physical markersreference genomephysical map• Gene syntenyTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Grouping the contigsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Assembly gapsphysical gapgroup Agroup Bsequencing gapssequencing gap - we know the order and orientation of the contigs and have atleast one clone spanning the gapphysical gap - no information known about the adjacent contigs, nor about the DNAspanning the gapTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Scaffolder outputPhysical gapsSequencing gapsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Problems with the data• Incorrect sizing of inserts– cut from gel – sizing is subjective– error increases with size• Chimeras (ends belong to different inserts)– biological reasons (esp. for large sized inserts)– sample tracking (human error)• Software must handle a certain error rate.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Ambiguous scaffoldI II IIIIIIIIIIIIIIII III I’ IITIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Unifying view of assemblyAssemblyScaffoldingTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

BAC-by-BAC Sequencing and ClosureWhy?• Aids complete closure by avoiding genomic repeatproblems• More streamlined collaboration with other sequencingand closure centers• Allows for data release and annotation of completedsequence throughout the projectWhy Not?• Complete closure not required• Small genome size• Time/cost concernsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

BAC-by-BAC Sequencing and Closure• Generate BAC-end sequence data• Choose and sequence “seed” BACs• Select minimally overlapping “tiling” BACs for contigextensionChromosome Marker1 Marker2 Marker3BAC LibraryMarker4Select BACs based on BAC-end overlaps, physical mapping data, etc.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Life of a BAC at TIGRIdentify BACBAC prep, fingerprintLibraryAssembleShotgunSequence test and verifyReactions, reassemblyCLOSED BACCompare digestsVerify BACTIGRTHE INSTITUTE FOR GENOMIC RESEARCHAnnotation

BAC Shotgun SequencingBAC DNASequencingShotgunLibraryEg. 4-6 KbAssembly and ScaffoldingContigsGap ClosureTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Genome Closure/FinishingTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Why Completeness is Important• Improves characterization of genome features– Gene order, replication origins• Better comparative genomics– Genome duplications, inversions• Determination of presence and absence of particulargenes and features is less subjective• Missing sequence might be important (e.g.,centromere)• Allows researchers to focus on biology not sequencing• Facilitates large scale correlation studies• Controls for contaminationTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

What Is Closure?• Obtaining sequence that was not obtained during randomsequencing which resulted in:• Sequencing Gaps• Physical Ends• Confirming the integrity of assemblies• Repeats and misassemblies• Verification of Clone Coverage• Confirming the base sequence of the consensus• Editing• Verification of Sequence CoverageTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Levels of Genome ClosureTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGR Finishing CriteriaA genome is considered finished and ready if it satisfies the followingcriteria:1. A continuous consensus of DNA sequence2. No ambiguous consensus basepairs3. At least 2X sequence coverage over the entire genome.a) both strands of one clone are sequencedb) two different clones sequencedc) Same clone sequenced with dual chemistry4. At least 2X clone coverage over the entire genome5. Complete confidence in all repetitive areasTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCHCauses for gaps• Non-random shotgun library– Toxicity of genes or promoters in E. coli– Genomic DNA difficult to clone (capsularpolysaccharides)– Unstable regions (low complexity)• Sequencing problems– Hard stops• Secondary structures• Very high or low GC content• Small unit tandem repeats– Loss of signal• Homopolymeric tracts• Very high or low GC content

Genome Closure StepsAssemble SequencesOrder and orient contigsValidate Sequence andAssembly IntegrityGap ClosureRepeat Identification andVerificationDesign PrimersSequencing GapsPhysical EndsTIGRTHE INSTITUTE FOR GENOMIC RESEARCHComplete BAC

Sequencing GapsContig AUnderlyingsequencesGAPPrimer 1 Primer 2Contig BTo close Sequencing gaps:A. Resequence short reads at contig endsB. Perform sequencing reactions:1. Design primers at contig ends2. Do sequencing reactions with:Insert 1 + primer 1 Insert 1 + primer 2Insert 2 + primer 1 Insert 2 + primer 2TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Physical EndsNo known links exist; there is no template available forsequencing.Contig CPhysical endTo link and close Physical ends use PCR:1. Design primers at contig ends2. If PCR products are obtained, sequence PCR products completelyTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Repetitive Areas• Repetitive areas are regions of high similarity within thegenome/BAC.• Sequences in these areas may be misassembled by theAssembler.• Verification of the sequence of repetitive areas:• A. Identify potential repetitive areas, using repeatFinder and other tools.• B. Classify repeats based on length, copy number, % similarity,structure and complexity.• C. If repeats are misassembled, transpose spanning clones or obtainPCR products and sequence to verify assembly.• Misassemblies can occur in two ways:TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

The base sequence of the repeat may bemisassembled:RPT 1C• Repeat is spanned by linking clones• Underlying sequences are unlinked clones, with mates inother assemblies (branched clones) or mates that did notsequence (dead end clones).• Results in incorrect consensus sequenceTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Mis-assembled repeatClones link different repeat flanksTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Identifying repetitive areasRepetitive areas are identified by the programrepeatFinderftp://ftp.tigr.org/pub/software/repeatFinder/• Analyzes sets of contigs for repeats greater than 50bp and then groups these repeats into classes• repeatFinder calls REPuter[Copyright© University of Bielefeld, Germanywww.genomes.de; Used with permission]TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence editor• In this example thereis an obviousdiscrepancy betweenthe base calls ofseveral of theunderlying clones inthis region.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Assembly Viewer• Several clones appear tobe misassembled• There are both obvioussize violations andorientation issuesTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Resultingcontig afterrepeatresolutionTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

AMOS Assembly InvestigatorToolbarPositionInsert andRead CoverageScaffoldFeaturesInsertsTIGRTHE INSTITUTE FOR GENOMIC RESEARCHCurrent Contig Position

Collapsed Repeat-5.5 CE DipReadCoverageSpikeIndividualCompressedMatesTIGRTHE INSTITUTE FOR GENOMIC RESEARCH68 Correlated SNPs

Resolved Repeat• Unique flank order is correct• Use linking information across the repeat (largeinsert clones or PCR)• Consensus sequence is correct• Use linked clones that have one mate in the repeatand the other anchored in unique sequence• Transposon mediated librariesTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence Validation: Editing• Finishers look at all areas where the consensussequence is below a certain quality threshold.• Electropherograms are used to determine which basecalls may need to be changed.• Only supporting sequences are edited, not theconsensus.• Editing provides an integrity check, may aid in assembly,identifying repeat areas and other unique problems.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence Validation:Sequence coverage1X2X3XSequence coverage rule:Every base in an assembly must be covered by at least twosequences of high quality.Why?Validating sequence coverage provides a high degree ofconfidence in the consensus base calls.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence Coverage in CloeTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Sequence Coverage in CloeTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Assembly Verification: CloneCoverageClone Coverage Rule:Every base must be spanned by two valid subclones.A valid subclone is a pair of forward/reverse mates that are:• Oriented correctly (pointing toward each other)• Separated by an approximate length of sequence determined bythe expected template size for the library.Why?Validating clone coverage provides a high degree ofconfidence that the contig was assembled correctly.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Clone Coverage ExampleCATCCGAAGACGATGCATCAATAGCATAClone-1 (F)Clone-1 (R)Clone-3 (R)Clone-2 (R)Clone-2 (F)Clone-5 (F)Clone-4 (R) Clone-6 (F) Clone-7 (R)0x1x• Red: 0 x clone coverage• Black: 1x clone coverage• Blue: 2x clone coverage2x1xTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Verify Closure by FingerprintComparisonTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Finished BAC Ready for Annotation!BAC DNASequencingShotgunLibraryEg. 4-6 KbContigsAssembly and ScaffoldingGap ClosureTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

EXTRA SLIDES STARTHERE…TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

HT TemplateLibraryConstructionHT SequencingClosure SequencingLIMSTrace ProcessingTrace ProcessingTrace DeliveryTrace DeliveryTrace ArchiveTIGRTHE INSTITUTE FOR GENOMIC RESEARCHProject DB

Representing assembliesconsensus startconsensus endconsensus ACTTTAGGAAGACGA-GAAGCCTGATAGGATCTCCGTCTCCGACaligned sequence CTGAC-ATGAA-CCTGATAGCAGsequence startsequence endAssembly table:consensus w/ and w/o gapsTIGRTHE INSTITUTE FOR GENOMIC RESEARCHAsmbl_link table:asmbl_idseq_nameseq_lend, seq_rendasm_lend, asm_rendlsequenceSequence table:sequenceend5, end3

Schematic View of 3 Typical ScaffoldsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Types of repeats• Short Tandem Repeats• Large Multi Unit Tandem Repeats• Multi Class Repeats• Inverted RepeatsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Short Tandem RepeatsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Large Multi Unit TandemRepeatsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Inverted RepeatsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Construction of Transposon-MediatedLibraries• Transposons are genetic elementsthat provide mobile primer siteswithin the target DNA forsequencing• The GPS-1 Genome PrimingSystem (New England BiolabsE7100S) enables transposoninsertions by using TnsABC*transposase+• Only one random insertion occursper target DNATIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Use of Transposon-mediated libraries:•Make a library using a transposon kit on a spanning clone.•Sequence the clones using the transposon primers•Assemble the sequencesPhysical map of thetransposon inserts:M13FM13FTransposon1 Transposon2N S N SRepeatClone DNAN1 S1M13RM13RN2S2TIGRFinal AssemblyTHE INSTITUTE FOR GENOMIC RESEARCH

Assembly of Transposon-Mediated Clone Walks12500 5000 7500 10000 1250015000GEGDK94TR2940 bpGEGDK94TFGEGDK94T175GEGDK94T48GEGDK94T22GEGDK94T6GEGDK94T34GEGDK94T73GEGDK94T72GEGDK94T36GEGDK94T41GEGDK94T77GEGDK94T106GEGDK94T10GEGDK94T12GEGDK94T47GEGDK94T189GEGDK94T90TIGRforward and reverse clone matestandem repeat regiontransposon insertionsiteTHE INSTITUTE FOR GENOMIC RESEARCH

Closure Challenges: SequencingThrough Secondary StructuresTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Hairpin structureTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Homopolymeric tractsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

• Apply different sequencing chemistries– Big-Dye terminator (default)– Dye-primer– dGTP mix (GC rich regions)• Denature structures - Additives– Betaine– DMSO• Break structure– Restriction digest– Transposon insertion– Micro-librariesTIGRTHE INSTITUTE FOR GENOMIC RESEARCHSolutions

Automated Finishing at TIGR• Closure is a feature driven activity• Feature Classification SystemTIGR– Identify/Classify Features• SEQ gaps• PHY gaps / scaffold ends• Low coverage• Repeat features• Feature Reaction Design Strategy– Bin by Reaction Type and Feature Size• Primer walks• PCR• Transposon BombingTHE INSTITUTE FOR GENOMIC RESEARCH

Automated Finishing (cont’d)• Laboratory Process Integration– Order reactions from the lab– Group reactions by lab process– Integrate changing laboratory automation– Monitor results• Reaction Product Resolution Strategy– Resolve reaction products back to features & contigs:conduct targeted assembly.– Evaluate resolution criteria for feature type.TIGRTHE INSTITUTE FOR GENOMIC RESEARCH

System ArchitectureCA PipelineInputProcessingFeatureIdentificationReactionDesignDBWork OrderPreparationReactionProcessingWork OrderManagementTIGRProductResolutionTHE INSTITUTE FOR GENOMIC RESEARCH

• Dead EndsBAC Project Challenges• Filling small gaps between adjacentsupercontigs• bp mismatches in BAC sequence overlaps• Project management/organizationTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

TIGRTHE INSTITUTE FOR GENOMIC RESEARCHDead EndsA Dead End occurs when a BAC at the end of asequence contig has no BAC-end hits or FPC BACsto extend the contig further• Design primers and PCR small unique region nearend of BAC• Hybridize to BAC filters to identify candidate BACs• End sequence and fingerprint candidate BACs toselect minimal tiling BAC• Sequence and Close new tiling BAC

Filling small gaps in a chromosome• GAP < 15kb:– PCR using genomic DNA template– Sequence and close PCR product• GAP > 15kb or unable to PCR:– Select gap-spanning BAC– Construct large insert library (10+kb inserts)– Sequence to low coverage (2-3x)– Scaffold mate pairs and select subclone tiling to close gapTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

p Mismatches In BAC OverlapsWhy?• Basecalling error in sequencing• Point mutation in BAC• Misassembly of repeats near BAC endWhat to do?• Examine sequence quality in each BAC• PCR region using genomic DNA template• Sequence PCR to verify correct basecallTIGRTHE INSTITUTE FOR GENOMIC RESEARCH

Project Organization• Database to store various BAC data:– Tiling, overlap info– Mapping, FPC info– Sequencing and Closure data– Collaborator’s BAC data• Web-based BAC status tracking andclosure task tracking toolsTIGRTHE INSTITUTE FOR GENOMIC RESEARCH