Genome Assembly and Finishing

Alla Lapidus, Ph.D.Associate Professor

Fox Chase Cancer Center

A typical Microbial (and not only) project

FINISHING

Annotation

Public release

Sequencing

Draftassembly

Goals:

Completely restore genome

Produce high quality consensus

Sequencing Technology at a Glance

Evolution of Microbial Drafts

Sanger only – 4x of 3kb plasmids + 4x of 8kb plasmids + 1x of fosmids – ~ $50k for 5MB genome draft

Hybrid Sanger/pyrosequence/Illumina – 4x 8kb Sanger + 15 x coverage 454 shotgun + 20x Illumina

(quality improvement)– ~ $35k for 5MB genome draft

454 + Solexa - 20x coverage 454 standard + 4x coverage 454 paired end (PE) + 50x

coverage Illumina shotgun (quality improvement; gaps) - ~ $10k per 5MB genome

Solexa only - low cost; too fragmented; good assembler is needed!

Solexa +PacBio - low cost; better sachffolding

Process Overview

Library Preparation - Sanger

DNA fragmentation

Random fragment DNA

Library Preparation - new

Assembly (assembler)

• Sanger reads only (phrap, PGA, Arachne) --

3kb-- --3kb--

--8kb-- --8kb--

---------40kb--------

• Hybrid Sanger/pyrosequence/Solexa (no special assemblers; use

Newbler, PGA, Arachne) 454 contig454 contig

--8kb-- --8kb-- --8kb--

--8kb-- --8kb-- --8kb--

454 shreds454 shreds

• 454/Solexa (Newbler, PCAP, Velvet, ALLPATH etc) –

Shotgun readsPE reads

Assembly: set of contigs

Draft assembly - what we get

10 16 21

10 21

Clone walk(Sanger lib)

Ordered sets of contigs (scaffolds)

New technologies: no clones to walk off even if you can scaffold contigs

(bPCR – new approach of gap closing)

16

PCR - sequence

pri1 pri2

PCR product

PE

Primer walking

PCR – sequence

(un captured gaps)

Template: gDNA

PCR product

Clone walk

(captured gaps)

Clone A

Why do we have gaps

• Sequencing coverage may not span all regions of the genome, thus producing gaps in the assembly – colony picking

• Assembly results of the shotgun reads may produce misassembled regions due to repetitive sequences (new and old tech)

• A biased base content (this can result in failure to be cloned, poor stability in the chosen host-vector system, or inability of the polymerase to reliably copy the sequence):

~ AT-rich DNA clones poorly in bacteria (cloning bias; promoters like structures {Sanger} )=> uncaptured gaps ~GC rich DNA is difficult to PCR and to sequence and often requires the use of special chemistry => captured gaps

~ high AT and GC content caused by problematic PCR (new tech)

What are gaps ?- Genome areas not covered by

random shotgun

Actual genome

Assembling repeats

High GC sequencing problems:

The presence of small hairpins (inverted repeat sequences) in theDNA that re anneal ether during sequencing or electrophoresisresulting in failed sequencing reactions or unreadable electrophoresisresults. (This can be aided by adding modifiers to the reaction,sequencing smaller clones and running gels at higher temperatures inthe presence of stronger denaturants).

Why more than one platform?

• 454 - high quality reliable skeletons of genomes (454 std + 454 PE): correctly assembled contigs; problems with repeats (unassembled or assembled in contigs outside of main scaffolds); homopolymer related frame shifts

• Illumina data is used to help improve the overall consensus quality, correct frameshifts and to close secondary structure related gaps; not ready for de-novo assembly of complex genomes (too many gaps!)

• Sanger – finishing reads; fosmids – larger repeats and templates for primer walk – less cost effective but very useful in many cases

454 (pyrosequence) and low GC

genomesThermotoga lettingae TMO

Sanger based draft assembly: - 55 total contigs; 41 contigs >2kb- 38GC% - biased Sanger libraries

Draft assembly +454- 2 total contigs; 1 contigs >2kb- 454 – no cloning

<166bp> - average length of gaps

454 and High GC projectsXylanimonas cellulosilytica DSM 15894 (3.8 MB; 72.1% GC)

PGA assembly - 9x of 8kb PGA assembly - 9x of 8kb +454

Assembly Total contigs Major contigs Scaffolds Misassenblies* N50

PGA-8kb 210 166 4 165 41,048

PGA-8kb+454 33 23 2 14 288,369

454/Sanger contig

Fosmid ends* and 454 PE

1.Pyrosequence and Sanger to obtain main ordered and oriented part of the assembly – Newbler assembler

3. Solexa reads to detect and correct errors in consensus –in house created tool (the Polisher) and close gaps (Velvet)

2. GapResolution (in house tool) to close some (up to 40%) gaps using unassembled 454 data – PGA or Newbler assemblers

Solexa

* Fosmids ends not used for microbes

Unassembled 454 reads

NextGen high Quality Drafts at JGI (multiple sequencing platforms)

Solexa contig

Solving gaps: gapResopution tool

ContigGap (due to repeat)

Read pairs that are found in contigs outside of this

scaffold

Step 1 For each gap, identify read pairs from contigs found on different scaffolds

Step 2 Assemble reads in contigs adjacent to the gap and reads obtained from contigs outside the scaffold. Sometimes use assembler other than Newbler for sub-assemblies (PGA)

Contig Gap

Consensus from sub-assembly

Solving gaps: gapResopution tool (II)

Step 3 If gap is not closed, tool designs designs primers for sequencing reactions

Contig Gap

Design sequencing reactions to close gap

Step 4 Iterate as necessary (in sub-assemblies)

http://www.jgi.doe.gov/degilbert@lbl.gov

• Velvet assembly• Blast Velvet contigs against Newbler ends• Use proper Velvet contigs to close gaps

Solexa for gaps

454 Contig GapVelvet contig

Illumina reads

Velvet contigs close gaps caused by hairpins and secondary structures

Low quality areas – areas of potential frameshifts

Assemblies contain low quality regions (red tags)

Frameshift 1 (AAAAA, should be AAAA)

Frameshift 2 (CCCC, should be CCC) homopolymers (n>=3)

Modified from N. Ivanova (JGI)

Homopoymer related frameshifts

Polisher: software for consensus software for consensus quality improvement quality improvement

Step 1: Align Illumina data to 454-only or Sanger/454 hybrid assembly

Contig

Illumina reads

Step 2: Analyze and correct consensus errorsC T

T

G

AAAAA

CorrectionsIllumina coverage >= 10X and at least 70% llumina bases disagrees with the reference base

Unsupporteda. Illumina coverage < 10Xb. Illumina coverage >= 10X and <70% of Illumina bases agree with the reference baseStep 3: Design sequencing reactions for low

quality and unsupported Illumina areas

Unsupported Illumina regionSanger/454 low quality

Errors corrected by Solexa

CCTCTTTGATGGAAATGATA**TCTTCGAGCATCGCCTC**GGGTTTTCCATACAGAGAACCTTTGATGATGAACCGGTTGAAGATCTGCGGGTCAAA CCTCTTTGATGGAAATAATA**TATTCGAGCATC TTAGTGGAAATGATA**TCTTCGAGCATCGCCTC CGAGCNTCGCCTC**GGGCTTTCCCT CGAGCATCGCCTC**GGGTTCTCCATACACAGA GCATCGCCTC**GGGTTTTCAATACAGAGAACCT CAGCGCCTC**GGGTTTTCCATACAGAGAACCTT ATCGCCTC**GGGTTTTCCAGACAGAGAACCTTT GGTTC**GGGTTTTCCATACAGAGAACCTTTGAT GTTTTCCATACAGAGAACATTTGATGATGAAC GTTGTCCATACAGAGAACTTTTGATGATGAAC TATANCATACAGAGAACCTTTGATGATGAACC ATTTCCAGACAGAGAACCNTTGATGATGAACC CAAACAGAGAACCTTTGAGGATGAACCGGTTG ACAGGGAACCTTAGATGATGAACCGGTTGAAG ACAGAGAACCTTAGATGATGAACCGGTTGAAG ACCGTTGATGATGAACCGGTTGAAGATCTGCG GATGGTGAACGGGTTGAAGATCTGCGGGTCAA GGTTTGAAGATCTGCGGGTCAAACCAGTCCTC GGTGGAAGATCTGCGGGTAAAACCAGTCCTCT GGT.GNAGAGCTGCGGGTCAAACCAGTCCTCTG TGAAGATCTGCGGTTCAAACCAGTCCTCTCCC GATCGGCGTGTCAAACCAGTCCTCTGCCTCGT TCTGCGGGTCAAACCAGTACTCTGCCTCGTTC

Frame shift detected (454 contig)

454 contig

Finished consensus

Sanger reads

So, what is Finishing?

The process of taking a rough draft assembly composed of

shotgun sequencing reads, identifying and resolving miss

assemblies, sequence gaps and regions of low quality to

produce a highly accurate finished DNA sequence.

Final error rate should be less than 1 per 50 Kb.

No gaps, no misassembled areas, no characters other than ACGT

Final quality:

Sequencing Centers for Archaea & BacteriaMay 2009: 3549 projects

JGI23%

JCVI18%

BROAD9%

WashU6%

BCM5%

WORLD37%

Genome projectsGenome projectsArchaea + Bacteria onlyArchaea + Bacteria only

http://www.genomesonline.org/

298CompleteGenomes

137CompleteGenomes

Metagenomic assembly and Finishing

• Typically size of metagenomic sequencing project is very large

• Different organisms have different coverage. Non-uniform sequence coverage results in significant under- and over-representation of certain community members

• Low coverage for the majority of organisms in highly complex communities leads to poor (if any) assemblies

• Chimerical contigs produced by co-assembly of sequencing reads originating from different species.

• Genome rearrangements and the presence of mobile genetic elements (phages, transposons) in closely related organisms further complicate assembly.

• No assemblers developed for metagenomic data sets

The whole-genome shotgun sequencing approach was used for a number of

microbial community projects, however useful quality control and assembly

of these data require reassessing methods developed to handle relatively

uniform sequences derived from isolate microbes.

QC: Annotation of poor quality sequence

To avoid this:

-make sure you use high quality sequence

-choose proper assembler

A Bioinformatician's Guide to Metagenomics . Microbiol Mol Biol Rev. 2008 December; 72(4): 557–578.

Assembly mistakes

A Bioinformatician's Guide to Metagenomics. Microbiol Mol Biol Rev. 2008 December; 72(4): 557–578.

Recommendations for metagenomic assembly

- Use Trimmer (Lucy etc) to treat reads PRIOR to assembly

- None of the existing assemblers designed for metagenomic data but assemblers like PGA work better with paired reads information and produce better assemblies.

- We currently test Newbler assembler for second generation sequencing: 454 only and 454/Solexa co-assembly

Metagenomic finishing: approach

Binning:Binning: Which DNA fragment

derived from which phylotype?

(BLAST; GC%; read depth)

Non-CAP readsNon-CAP reads

CAP readsCAP reads

++

Complete genome of Complete genome of Candidatus Accumulibacter

phosphatis

Lucy/PGALucy/PGA

Candidatus Accumulibacter phosphatis (CAP)

~ 45%

Few more details: read quality

Merged assemblies ( k=31 and k=51) with minimus(Cloneview used for visualization)

Green k=31

Purple k=51Illumina only data

Stats for 31, 51 and merged 31-51 assemblies

Hash L 31 51 31_51expCov NO NO NOTotal Ctgs 3,796,782 377,044 275,273Largest 15,553 23,012 40,135N50bp 116 196 325Min Ctg L 80 80 80Total Len Ctgs 360,994,462 62,631,932 138,833,812

Thank you!