1. Next-Generation Sequencing of Microbial Genomes and Metagenomes
Christine King Farncombe Metagenomics Facility
Human Microbiome Journal Club
July 13, 2012

2. Overview Next-generation sequencing Project overview Applications
Instruments
Library prep and sequencing chemistry
Sequence quality
Project overview
Microbial genomes
Microbial communities

3. DNA Sequencing 1st generation 2nd generation (NGS) 3rd generation
Sanger chain termination
Capillary electrophoresis
2nd generation (NGS)
High throughput, “massively parallel”
Shorter reads
Sequencing-by-synthesis
3rd generation
Single molecule
Nanopores

4. Applications DNA sequencing RNA sequencing De novo genomes
Resequencing
Shotgun (e.g. mutant strains)
Amplicon (e.g. HLA, cancer)
Sequence capture (e.g. exome)
Metagenome
Amplicon (e.g. 16S, COI, viral)
Shotgun
ChIP
RNA sequencing
Gene expression
Gene annotation, splice variants
Metatranscriptome

5. Instruments

6. Instruments Instrument # of reads Read length (bp) Total output (Gb)
Cost per base
Run Time
Technology
GS FLX 1M
450
0.5
$$$$ ++
emPCR, SBS, light detection
GS FLX+
650
0.6
GS Jr
100K
0.05
GAIIx
640M
2x 150 90 $$
+++
Bridge PCR, SBS, fluororphore
HiSeq 2000 6B
2x 100
600 $
MiSeq
12M 2
PacBio RS
>10K
>1000
0.01 +
Single-molecule seq, fluorophore
SOLiD 5500xl
1.4B
155
emPCR, probe ligation, fluorophore
Ion PGM - 316
>100
0.1
$$$
emPCR, SBS, pH change
Ion PGM - 318 6M 1

7. Which instrument(s) to use?
Read length vs number of reads
Cost per base, per sample, per project (multiplexing?)
Accuracy
Run time, wait time
Application
Length
# Reads
Accuracy
Instruments
Considerations
De novo (small)
+++ ++
MiSeq, 454, Ion
Mix lengths
De novo (large)
HiSeq, 454, SOLiD
Mix lengths, MP
Re-seq (small)
MiSeq, Ion
Multiplex?
Re-seq (large)
HiSeq, SOLiD
Enrichment?
RNA-seq (count) +
Illumina, SOLiD, Ion
Ref? Size? Rare?
Amplicons
454, MiSeq
Size? Multiplex?
Metagenomics
Illumina, 454, SOLiD
Length vs depth

8. Library Preparation
Goal: fragments of DNA, each end flanked by adaptor sequences
Adaptors contain amplification- and sequencing primer binding sites; platform- and chemistry-specific
Optional: sample-specific barcodes/indexes/MIDs/tags allow multiplexing during sequencing
Library QC: quantity, size

9. Library Preparation Library types: Shotgun (DNA) Mate pair (DNA)
May begin with ChIP
May follow with sequence capture
Mate pair (DNA)
Amplicon (DNA)
Total RNA
May enrich for mRNA (poly-A enrichment, rRNA depletion)
Convert to cDNA (then similar to DNA protocols)
Small RNA
RNA ligations, convert to cDNA after

10. Library Preparation: Shotgun
Fragmentation
Sonication
Nebulization
Enzymatic
End repair
3’ overhangs digested
5’ overhangs filled
5’ phosphate added

11. Library Preparation: Shotgun
Adapter ligation
T-overhangs
Forked structure controls orientation
Library amplification
Few cycles
Enrich for correctly-adapted fragments
Required to complete adapter structure in some protocols
Size selection
Gel excision, AMPure beads
Limit insert size as needed, remove artifacts

12. Library Preparation: Amplicon
Amplify region of interest using PCR
Primers contain adapter sequences

13. Library Preparation: Mate Pair
Begin with large fragments (e.g. 3kb, 20kb)
Circularize and fragment again
Illumina: direct ligation
454: Cre/Lox recombination
Enrich for fragments containing the junction
Proceed with shotgun library prep

14. Library Preparation: Mate Pair
Why? Paired sequences are a known distance apart; improves genome assembly
Note: 454 calls these “paired end libraries”, not to be confused with Illumina’s “paired end sequencing”!

15. Sequencing: Illumina Cluster generation ~800K clusters/mm^2
Library fragments hybridize to oligos on the flow cell
New strand synthesized, original denatured, removed
Free end binds to adjacent oligos (bridge formation)
Complimentary strand synthesized, denatured (both tethered to flow cell)
Repeat to form clonal cluster
Cleave one oligo, denature to leave ssDNA clusters
~800K clusters/mm^2

16. Sequencing: Illumina Variety of workflows: Single- or paired end reads
0, 1, or 2 index reads

17. Sequencing: Illumina
At each cycle, all 4 fluorescently-labeled nucleotides pass over the flow cell
Each cluster incorporates one nt (terminator) per cycle
Fluor is imaged, then cleaved
De-block and repeat

18. Sequencing: Illumina Other terminology: File format: fastq
cBot – accessory instrument that performs cluster generation
Lanes – divisions (8) of HiSeq and GAIIx flow cells
PhiX – bacteriophage with small, balanced genome; PhiX library spiked in with samples for QC
Phasing/pre-phasing – nt incorporation falls behind or jumps ahead on a portion of strands in the cluster and contributes to noise
Chastity filter – measures signal purity (after intensity corrections); if the background signal is high, cluster will be discarded
BaseSpace – cloud computing site for processing MiSeq data
File format: fastq

19. Sequencing: 454
emPCR: clonal amplification of bead- bound library in microdroplets
Library input amounts critical!
One molecule per bead
Titration procedure

20. Sequencing: 454 Library capture: beads coated with complimentary oligo
Amplification: droplet contains PCR reagents and the other oligo
Post-PCR: millions of identical fragments attached to the bead

21. Sequencing: 454 Bead Recovery: physical and chemical disruption
Enrichment: capture successfully amplified beads using biotinylated primers + magnetic, streptavidin beads

22. Sequencing: 454 Deposit bead layers onto PicoTiterPlate: Enzyme beads
Enriched DNA beads
More enzyme beads
PPiase beads

23. Sequencing: 454

24. Sequencing: 454 Pyrosequencing 4 nucleotides flow separately
If nt incorporation…PPi...light
APS + PPi (sulfurylase)  ATP
Luciferin + ATP (luciferase)  light + oxyluciferin
Amount of light proportional to #nt incorporated
Rinse and repeat with next nt

25. Sequencing: 454
Camera captures light emitted from every well during every nucleotide flow

26. Sequencing: 454
Flowgram: representation of a sequence, based on the pattern of light emitted from a single well

27. Sequencing: 454 Other terminology:
Lib-L/Lib-A: adapter variants, “ligated” or “annealed”
Titanium chemistry: ~450 bp reads on all instruments
XL+ chemistry: ~700 bp reads on the FLX+ instrument
Flow: one of the four nucleotides flows over the PTP
Cycle: a set of four flows, in order
Valley flow: if number of bases incorporated in a given read during that flow is uncertain, e.g. 1.5 units of light (background signal, homopolymers)
File format: sff (standard flowgram format)

28. Sequencing: Ion Torrent
Procedures and chemistry similar to 454
Instead of PPi, measure H+ release (pH change) via semiconductor chip
No expensive camera or laser required, no modified nucleotides

29. Probability of Error (P)
Sequence Quality
Error probabilities determined using training sets, platform- specific biases
Expressed as a quality value (QV or Q score) per base
Similar to PHRED scores:
Q = -10 log10P
P = 10 -Q/10
Phred (Q) Score
Probability of Error (P)
Base Call Accuracy 10
1 in 10
90% 20
1 in 100
99% 30
1 in 1K
99.9% 40
1 in 10K
99.99% 50
1 in 100K
99.999%

30. Project 1: Microbial Genome
Considerations:
Reference genome?
How much coverage do I want?
How big is the genome
How much data do I need?
bp needed = genome size X coverage
Which instrument/chemistry configuration to use?
Coverage
Depth (number of times a particular base is “covered” by a read (e.g. 25X)
Breadth (% of genome with at least 1X coverage)

31. Project 1: Microbial Genome
Sample preparation
Isolate high quality (not degraded) and high purity (no RNA) gDNA
Verify on a gel
Quantify using dsDNA-specific dye
Library preparation
Can do this yourself if you like
~ $200 per sample for Nextera
Cheaper protocols
Cheaper in bulk
Barcode compatibility

32. Project 1: Microbial Genome
Library QC
Insert size confirmed on BioAnalyzer (within range, no artifacts)
Pool barcoded libraries (normalize based on PicoGreen quantification)
Absolute quantification of library pools using qPCR

33. Project 1: Microbial Genome
MiSeq sequencing
Dilute and denature library pool (optimal concentration requires titration...)
Spike in PhiX library as needed (e.g. 1%)
Prepare and load reagents, flow cell
Basic filtering and de-multiplexing performed automatically
Download fastq files from BaseSpace

34. Project 1: Microbial Genome
Data processing
Additional filtering
Trim the ends
Remove PCR duplicates
Assembly: overlapping reads are assembled to eachother based on sequence similarity = contigs

35. Project 1: Microbial Genome
What’s next?
Polish the genome (hybrid assemblies, mate pair libraries)
Annotate (ORFs, RNA- seq)
Compare

36. Project 2: Microbial Community
Shotgun metagenomics
Unbiased survey of community content
Random library fragments may provide very little taxonomic resolution (e.g. conserved, unknown)
Identify genes, classify by function
Targeted metagenomics
Limited survey of community content
Targeted loci provide excellent taxonomic resolution, but may exclude certain taxa
Identify OTUs, classify by taxonomy

37. Project 2: Microbial Community
16S rRNA
Multi-copy gene (1.5 kb)
Conserved and hypervariable regions
Extensive databases from known species

38. Project 2: Microbial Community
Considerations:
Biases in sampling methods, culturing, DNA isolation, PCR...replicate
Available SOPs
How many reads per sample?
Read length matters!
Sample preparation:
Isolate DNA
PCR amplify, purify
High-fidelity polymerase
Barcoded primers
No primer dimers!
Normalize PCR products and pool

39. Project 2: Microbial Community
454 Sequencing
emPCR titrations with different library input
Bulk emPCR
Sequence
Basic filtering
Collect sff files
Data processing
De-multiplexing
Additional filtering
Trim the barcodes, primers
Check for chimeras

40. Project 2: Microbial Community
Clustering
Sequences grouped by similarity = OTUs

41. Project 2: Microbial Community
Taxonomic identification
OTUs are classifed by comparing to known 16S sequences
Level of classification (e.g. family vs genus)?
Diversity
Within sample
Between samples