1. Single Cell, RNA, & Chromosome Sequencing Technologies
George Church 2:30- 3:00 PM Tue 3-Oct-2006
Cancer Genomics & Emerging Technologies
Thanks to: NCI/NIH HMS-CGCC
AppliedBiosystems-Agencourt,
Affymetrix, Helicos, 454, Solexa, DNAdirect, CompleteGenomics, Codon Devices

2. Muliplex Polony Summary
Technologies for selecting genomic regions
Mbp scale for rearrangements
RNA tags & spliceforms
1 to 200 bp scale for SNPs & exons (1%)
Low cost & high accuracy : $.07/kbp at 3E-7 errors
Paired-end-tags (PET) for rearrangements
Detection of rare mutations (e.g. drug resistance alleles)
60 million reads per run

3. Selective genome sequencing
Numerous (100K) Small Regions (exons & point mutations)
PCR : 21 Mbp >$250K Sjoblom et al (2006) Science
Highly multiplexed molecular inversion probe genotyping:
over 10,000 targeted SNPs genotyped in a single tube assay.
Hardenbol et al. Genome Res Feb;15(2):
Analyzing genes using closing and replicating circles.
Nilsson et al. (2006) Trends Biotechnol 24:83.
One large region
Single molecule amplification 1 to 4 Mbp
Zhang et al Nature Biotech. 24:680
Direct genomic [BAC hybridization] selection. [50% pure]
Bashiardes et al (2005) Nat Methods 2: 63.

4. Selective genome sequencing
Two ways to capture alleles from
genomic ss-DNA
In vitro
Paired-tag
library
Gap
fill
Cleave
& ligate
Red=Synthetic; Yellow=genomic
Shendure, et al. Science 309(5741):
Nilsson et al. (2006) Trends Biotechnol 24:83.
How do we optimize >100K 100mers ?
Zhang, Chou, Shendure, Li, Leproust, Church, Dahl, Davis, Nilsson

5. How? 10 Mbp of oligos / $1000 chip
~1000X lower oligo costs
Digital Micromirror Array
8K Atactic/Xeotron/Invitrogen
Photo-Generated Acid
12K Combimatrix/Codon Electrolytic
44K Agilent Ink-jet standard reagents
380K Nimblegen/GA Photolabile 5'protection
Amplify pools of 50mers using flanking universal PCR primers &
3 paths to 10X error correction
Tian et al. Nature. 432:1050; Carr & Jacobson 2004 NAR; Smith & Modrich 1997 PNAS

6. Padlock, Molecular Inversion Probes (MIPs)
CG to CA,TG 35% of germline, 44% of colorectal cancer mutations
(not restricted to single nucleotides nor common polymorphisms) R
Universal primers
Optional multiplex tag L
Genomic DNA CG CA TG
Alternative alleles
Zhang, Chou, Shendure, Li, Leproust, Church, Dahl, Davis, Nilsson
(10K to 1M 100-mer probes per pool -- see Kun Zhang’s poster)
Vitkup, Sander, Church The Amino-acid Mutational Spectrum of Human Genetic Disease. Genome Biol. 4: R72. (CG to CA, TG)

7. Sequencing genomes from single cells via polymerase clones -- Plones
(single chromosome, cell , RNA or particle)
Zhang, et al. (2006) . Nature Biotech. June ’06
1) When we only have one cell as in Preimplantation
Genetic Diagnosis/Haplotyping (PGD/PGH)
or environmental samples (poor lab growth)
2) Candidate chromosome region sequencing
3) Prioritizing or pooling (rare) species based
on an initial DNA screen (metagenomics)
4) Multiple chromosomes in a cell or virus
5) RNA splicing
6) Cell-cell interactions
(predator-prey, symbionts, commensals, parasites)
Announce speaker names; recognize Jim & George
Phi-29 Polymerase Stand-displacement amplification

8. Single molecule amplification sequencing
Multiple Displacement Amplification (MDA)
NBT (2006) 24:
Note!: Single human cell 1000X easier than 5 Mbp
Zhang et al., Nature Biotechnology (2006) 24:680

9. Single-cell sequencing: 4.7 Mbp (plones)
Ultra-clean conditions for reduction of background amplification + Real-Time monitoring
Post-amplification chip hybridization distinguishes alleles
Amplification variation random & easily filled by PCR

10. CD44 Counts (RNA splicing forms)
Eph4 = mammary epithelial cell line
Eph4bDD = stable transfection of Eph4 with MEK-1 (tumorigenic)
Zhu, Shendure, Mitra, Church, Single Molecule Profiling of Alternative Pre-mRNA Splicing. Science 301:836-8.

11. Beads or not, Ligase or Polymerase
Reading Polonies
Beads or not, Ligase or Polymerase A G C T

12. ‘Next Generation’ Sequencing Status
Multi-molecule Reaction Volume
AB/APG Ligase beads 1 fL
454/Roche Pol beads ,000 fL
Solexa Pol term fL
CGI Ligase fL
Affymetrix Hybr array fL
Single molecules
Helicos Biosci Pol <1fL
Visigen Biotech Pol FRET <1fL
Pacific Biosci Pol <1fL
Agilent Nanopores <1fL
fL =1E-15 liters
(femto)
(7/9 involve our lab)

13. Length& run-time vs. Accuracy&Cost
"Future improvements in the read lengths, demonstrated at 7 consecutive bases per tag (Shendure et al., 2005) and reductions in the run time, currently 60 hours, will make this a useful platform for resequencing."
--Leamon, et al. (454) Gene Therapy and Regulation 3: 15-31 
Note that without ‘future improvements’:
Affymetrix/Illumina read-lengths of 1 base per tag are useful.
60 million reads/run is 10X faster per read than 500K reads/run.
& 50X lower cost per bp due to lower reagent & instrument costs.
$500/run
$140K

14. Polony Sequencing Equipment
CCD camera
microscope with
xyz controls
Autosampler
(96 wells)
(HPLC-like)
flow-cell
syringe pump
temperature
control

15. Monolayer immobilization
Integrated Polony Sequencing Pipeline (open source hardware, software, wetware)
Monolayer immobilization
In vitro
paired tag
libraries
Bead polonies via
emulsion PCR
Enrich amplified beads
Dressman et al PNAS 2003
SBE or SBL
sequencing
SOFTWARE
Images → Tag Sequences
Tag Sequences → Genome
Epifluorescence & Flow Cell $140K
Shendure, Porreca, Reppas, Lin, McCutcheon, Rosenbaum, Wang, Zhang, Mitra, Church (2005) Science 309:1728.

16. 4 positions for paired-end anchor 'primers'
ePCR bead L
Tag 1 M
Tag 2 R 5’ 3’
7 bp
6 bp
7 bp
6 bp
4 positions for paired-end anchor 'primers'
Each yields 6 to 7 bp of contiguous sequence
26 bp new sequence per 135 bp amplicon

17. Sequencing by Ligation (SBL) with fluorescent combinatorial 9-mers
Excitation Emission
5’-Cy5-nnnnAnnnn-3’
5’-Cy3-nnnnGnnnn-3’
5’-TR-nnnnCnnnn-3’
5’-Cy3+Cy5-nnnnTnnnn-3’ nm
5'PO4
ACUCAUC…
(3’)…TAGAGT????????????????TGAGTAG…(5’)
Shendure, Porreca, et al. (2005) Science 309:1728

18. Why low error rates?
Goal of genotyping & resequencing  Discovery of variants
e.g. cancer somatic mutations 4E-6 (&lab-evolved cells)
Consensus error rate Total errors (E.coli) (Human)
1E-4 Bermuda/Hapmap ,000
4E ,000 3E
1E-8 Goal for
Also, effectively reduce (sub)genome target size by enrichment for exons or common SNPs to reduce cost & # false positives.

19. Microbial lab evolution
Lenski Citrate utilization
Church Trp/Tyr exchange
Palsson Glycerol utilization
Edwards Radiation resistance
Ingram Lactate production
Stephanopoulos Ethanol resistance
Marliere Thermotolerance
J&J Diarylquinoline resistance (TB)
DuPont 1,3-propanediol production

20. Polony-based Whole-Genome Mutation Discovery of DTrp clone
Position
Type
Gene
Location
Function
Mechanism
986,334
T > G
ompF
Promoter-10
Promoter of Non-specific transport channel
Makes promoter more consensus-like
985,797
Glu > Ala
Non-specific transport channel
Makes pore bigger and more hydrophobic
931,960
D8 bp
lrp
frameshift
General Transcriptional Regulator ?
ompF – non specific transport channel
Glu-117 → Ala (in the pore)
Charged residue known to affect pore size and selectivity
Can increase import & export capability simultaneously
Shendure, et al. (2005) Science 309:1728

21. Multiple Genotypes, Similar Themes
Evolving Population:
Multiple Genotypes, Similar Themes
PCR amplification and sequencing of OmpF and Lrp from multiple clones from 3 independent lines of Trp/Tyr co-cultures:
OmpF: 42R G, L, C, DV, 117 EA
Arg  Gly, Leu, Cys ; Asp  Val; Glu  Ala
Hydrophillic and bulky  hydrophobic and smaller
Promoter: -12AC, -35 CA
More consensus like
Lrp: 1bp deletion, 9bp deletion, 8bp deletion, IS2 insertion,
R->L in DBD.
Change in global gene regulation?
Heterogeneity within each time-point reflects colony heterogeneity.
Reppas, Lin, et al (unpublished)

22. Mixture of wild & 2kb Inversion (pin)
proximal tag
placement
Incorrect distance
Red=same strand
Black opposite strand
distal tag
placement
1,206k
1,210k
Using paired ends, rearrangement & copy-number detection is >1000X easier than point mutation detection (6X vs 6000X)

23. Polonies for human inversions
>300 kbp long inverted repeats
Turner, Hurles, et al Nat Methods 3: Sanger Inst. & HMS

24. Polonies for haplotyping, recombination, LOH
Sequencing/genotyping on single human chromosomes
Polonies for haplotyping, recombination, LOH
153 Mbp
Zhang et al. Nature Genet. Mar 2006

25. Monitoring resistance to BCR-ABL-kinase inhibitors with polonies during CML patient therapy Nardi, Raz, Chao, Wu, Stone, Cortes, Deininger, Church, Zhu, Daley (submitted)
M244V
T315I
E255K

26. Muliplex Polony Summary
Technologies for selecting genomic regions
Mbp scale for rearrangements
RNA tags & spliceforms
1 to 200 bp scale for SNPs & exons (1%)
Low cost & high accuracy : $.07/kbp at 3E-7 errors
Paired-end-tags (PET) for rearrangements
Detection of rare mutations (e.g. drug resistance alleles)
60 million reads per run

27. .

28. .

29. Polonies with & without beads or gels
Increases from 14 to 57 million polony beads per run & improves data quality.
Kim, Porreca, Seidman, Church
unpublished

30. Why low error rates?
Goal of genotyping & resequencing  Discovery of variants
e.g. cancer somatic mutations 4E-6 (&lab-evolved cells)
Consensus error rate Total errors (E.coli) (Human)
1E-4 Bermuda/Hapmap ,000
4E ,000 3E
1E-8 Goal for
Also, effectively reduce (sub)genome target size by enrichment for exons or common SNPs to reduce cost & # false positives.

31. Cost vs consensus error rate
Polony Sep05 Sep 06
AB Sep05
$ $
M K $30K
Paired ends yes no yes
Device $ K K K
Announce speaker names; recognize Jim & George
Cost vs consensus error rate

32. Cancer exon sequencing
$250K per sample (13,023 genes, 21 Mbp, 135,483 primer pairs)
using PCR & capillary sequencing.
$3K per sample (estimate) using single tube capture & polonies
Sjoblom et al. The Consensus Coding Sequences of Human Breast and Colorectal Cancers. Science Sep;
Davies et al. Somatic mutations of the protein kinase gene family in human lung cancer. Cancer Res :