1. DARPA BAA 01-26: BIO-COMP http://arep.med.harvard.edu http://arep.med.harvard.edu Technical challenges and risks: “DNA computing” so far focused on computing (which is slow relative to silicon computing). This project, in contrast, explores one of the few working sub- nanometer input/output, memory & manufacturing processes. The major challenge is integration with silicon computing while maintaining the nano-size advantages. Church Lab: DNA memory I/O

2. 85% Single DNA molecule detection as amplified “polonies”

3. Polony Template 3’ P’ P 5’ AATACAATTCACACAGGAAACAGCTATGACATTC TATTGTTAAAGTGTGTCCTTTGTCGATACTGGTA…5’ FITC ( C )CY3 ( T ) Mean Intensity: 58, 0.5 40, 6.5 0.3, 48 0.4, 43 Proof of 26 steps of 0.34 nm each

4. Future of computing & unattended bio-monitoring 3D memory arrays sub-nm 3 per bit (currently 10 14 nm 3 ) Input: light, chemicals/toxins Output: nm-scale positioning Church Lab: DNA memory I/O DoD impact

5. Church Lab: DNA memory I/O Deliverables (full version page 1 of 2) 1 DNA memory: 1#1: Find SSR upper-limit & oligonucleotides required for reversible, irreversible, & interwoven systems 1#2: Establish alternative to Flp involving Lambda-Red homologous recombination. 1#3: Design a variant usable in whole organism cell-lineage monitoring. 1#4: We will test an error-prone polymerase in year two with DNA fluorescent-base extension as the output. 1#5: A goal of 2 kilobits of storage at one bit-per-base in a year 3-4 option. 2 A/D Input: in vivo allostery, in vitro light 2#1: TetR or other sensor inputs will be studied and integrated into a recombinational system. 2#2: Effects of DNA length on Zinc finger specificity, 5 domains harvested from a genome. 2#3: Insert the cAMP-triggered ribozyme at the 5' end of GFP such cAMP affects translation. 2#4: Incorporation of a fluorescent nucleotide by RNA polymerase under control of a caged ATP or GTP. 2#5: Use these to regulate the incorporation of dNTPs in the format of an RNAPol-etaDNAPol fusion. 3 D/A Output: polonies & micromirrors 3#1: 3D arrays of DNAs & polymerases will use in situ sequencing in gels & chemically cleavable dNTPs. 3#2: Photocleavable versions of the above will be tested. 3#3: A microfluidic synthesis of 3000 oligomers (30-mers) will be synthesized using a micromirror array 3#4 Bridge a one-cm pattern with nm precision using a chemi-optical combination. Controlled polymerase stepper-positioners will incorporate branching oligonucleotides at nm scale. 3#5 Incorporation of electronic or optical computing components will be attempted (option)

6. Deliverables (page 2 of 2) In vitro minigenome: 4#1: A synthetic genome replication dependent in vitro translation of GFP or luciferase using E.coli lysates. 4#2: As above but showing dependence on added EF-Tu of measures of replication rate. 4#3: A highly purified translation system will be used to extend the results of 4#1 or 4#2. 4#4: A SSR or recombination method will be tested for the transfer of genes for expression in minigenomes. 4#5: We will develop a rolling circle primer immobilization method. 4#6: We will assess the doubling rates varying number of primers, or amount of nicking. Computational: 5#1: Coordination with other DARPA BIO-COMP participants. 5#2: The known 3D models will be adapted to include any sequence changes initially EF-Tu mutations. 5#3: Software for inclusion of microscopic morphology, dynamics, and stochastics will be developed. 5#4: As an alternative to the direct allosteric design, we will develop a model for signal transduction cascade, which will be based on two-hybrid and microarray data 5#5: Metrics for optimality and quality assessment and outlier identification will be developed and refined.