Ecologies, Economies, & Exponentials in Modeling Technological Goals & Human Impact

1. Ecologies, Economies, & Exponentials in Modeling Technological Goals & Human Impact George Church The Future of Human Nature: Promises and Challenges of Revolutions in Genomics and Computer Science Boston University 12-Apr-2003

2. In possibly less than 250 years ? • Our doctor’s actions depend on knowing whether we & our viruses/bacteria are drug sensisitive or resistant. • Cars, bikes & pedestrians have their own spaces. • Hackers have less access to our medical records • & doctors more access.

3. Future of Human Nature Systems Biology & Modeling Speculative Specifications Potential timelines Limits to Mass Limits to Cost (energy & environment) How? Potential pathways to nanotechnology Human biology keeping up: stem cells

4. gggatttagctcagttgggagagcgccagactgaa gat ttg gag gtcctgtgttcgatccacagaattcgcacca Models:Share, Search, Merge, Check, Design

5. Systems Biology Models Environment Metabolites RNAi Insertions SNPs DNA Proteins RNA Replication rate interactions Ecosystems Cancer & stem cells Darwinian optima Molecular machines

6. Compare predicted with observed protein properties (abundance, localization, postsynthetic modifications)E.coli Link et al. 1997 Electrophoresis 18:1259-313 (Pub)

7. Compare Flux data with predicted optimaSegre et alPNAS (2002)99:15112-7 200 WT (LP) 180 7 8 160 140 9 120 10 Predicted Fluxes r=0.91 p=8e-8 100 11 14 13 12 3 1 80 60 40 16 20 2 6 5 15 4 17 18 0 0 50 100 150 200 Experimental Fluxes 250 250 Dpyk (LP) Dpyk (QP) 200 200 18 7 r=0.56 p=7e-3 8 r=-0.06 p=6e-1 150 150 7 8 2 Predicted Fluxes Predicted Fluxes 10 9 13 100 9 100 11 12 3 1 14 10 14 13 11 12 3 50 50 5 6 4 16 16 2 15 5 6 18 17 15 17 0 0 4 1 -50 -50 -50 0 50 100 150 200 250 -50 0 50 100 150 200 250 Experimental Fluxes Experimental Fluxes

8. Genetically modified organisms

9. Limits to diversity?

10. Humans past & future? Past Future Locomotion 50 26720 km/h Ocean depth 75m 4500m Visible l .4-.7 m pm-Mm Cold 0oC 3oK Memory 20 yr 2000 yr

11. Human Nature? 7 deadly sins: Pride, Envy, Greed, Lust, Gluttony, Sloth, Anger + 7 heavenly virtues: Faith, Hope, Charity, Fortitude, Justice, Temperance, Prudence +1 Darwinian breakthrough: Hyperadaptability via anticipatory modeling & experimentation

12. Inheritance is not just DNA

13. Is germline engineering the fastest threat/promise? Watson-Crick base pair (Nature April 25, 1953)

14. Somatic vs germline engineering • Days from design to phenotype vs 20 years • Ethics of adults choosing for children • Phenotype is more “predictable” from adult than embryo • Histocompatible adult stem cells may be more accessible • Interfaces with inorganic engineering

15. Steeper than exponential growth If population growth decreases with increasing GDP, then #CPUs could overtake #people.

16. Steeper than exponential growth http://www.faughnan.com/poverty.html

17. Vertebratebrain size evolution • Harry J. Jerison, Paleoneurology & the Evolution of Mind, Scientific American, Jan 1976 http://serendip.brynmawr.edu/bb/brainevolution/brainevol3.html

18. The human neural net fig “The retina's 10 million detections per second [.02 g] ... extrapolation ... 1014 instructions per second to emulate the 1,500 gram human brain. ... thirty more years at the present pace would close the millionfold gap.” (Morovec99) Edge & motion detection (examples)

19. Steeper than exponential growth Kurzweil/Moore's law (Instructions per second/ $K) http://www.kurzweilai.net/meme/frame.html?main=/articles/art0184.html

20. Computing limits Current: 1010 IPS/kg on 1010 bits Limit: 1051 IPS/kg on 1031 bits Lloyd Nat 406:1047

21. Computing costs • High parallelism & density: DNA: 1 nm3 /bit ; CD: 1013 nm3 /bit • DNA copying near thermodynamic limit • 2 x1019 op/J for DNA copying • 34 x1019 op/J thermodynamic limit • 109 op/J for conventional computers • 1011 op/J for human brain @ 1kW

22. Global Economies Launching a satellite: $0.3 billion http://electronics.howstuffworks.com/satellite8.htm Eradication of polio by 2006 by vaccine $0.3 billion http://www.abpi.org.uk/publications/publication_details/prevention/section9.asp Sequencing a human genome: $0.3 billion http://www.nih.gov/news/pr/mar2003/nhgri-04.htm Costs of NOT eradicating polio : $3 billion Bloom: www.ellison-med-fn.org/files/WoodsHole.ppt Government change for 22M people: $80 billion www.cnn.com Hackers & e-viruses $1.6 trillion (5% global GDP) http://www.vnunet.com/News/1106282

23. CO2 100 ppmv increase http://jan.ucc.nau.edu/~doetqp-p/courses/env470/Lectures/lec41/Lec41.htm

24. Global Energy & CO2 Fluxes 40-170 x1015 W of sunlight hits earth. We consume 1-2 x 1013 W per 6x109 people. (2kW each). CO2 >370 ppm = 730 x1015 g globally, increase ~3 x1015 /yr. Ocean productivity = ~100 x1015 g/yr. Autotrophs: 1025 Prochlorococcus cells globally (108 per liter) Undone by Cyanophages & Heterotrophs: 2x1028 SAR11 cells in the oceans http://www.gsfc.nasa.gov/gsfc/service/gallery/fact_sheets/earthsci/terra/earths_energy_balance.htm http://www.poemsinc.org/factsenergy.html Morris et al. Nature 2002 Dec 19-26;420(6917):806-10. http://hosting.uaa.alaska.edu/mhines/biol468/pages/carbon.html

25. Limits to mass Lithosphere (0.2% C, 75% SiO2) 110 C at 4 km Diameter = 1.3e6 m = 5e22 g (5000 species / g soil) Microbial hydrosphere 1.4e21g = 1e27 cells = 4e33 bp Biosphere 3e15 g (dry wt. marine); 2e18 g (land) Anthrosphere (23% C) 6e14g = 6e23 cells = 4e32 bp. fig

26. Potential economically viable and humane pathways Avoiding “Colossus” and “grey goo” To maintain diversity some of us may need to escape Mars 3 years 1 Sv (0.4 Sv ISS, 2 mSv earth, 0.02 mSv/Xray, Deinococcus 20kSv)

27. Steeper than exponential growth http://www.kurzweilai.net/meme/frame.html?main=/articles/art0184.html

28. Why improve measurements? Human genomes (6 billion)2 = 1019 bp Immune & cancer genome changes >1010 bp per time point RNA ends & splicing: in situ 1012 bits/mm3 Biodiversity: Environmental & lab evolution Compact storage 105 now to 1017 bits/ mm3 eventually & How? ($1K per genome, 108-1013 bits/$ ) • The issue is not speed, but integration. • Cost per 99.99% bp : Including Reagents, Personnel, • Equipment/5yr, Overhead/sq.m • Sub-mm scale : 1mm = femtoliter (10-15) • Instruments $2-50K per CPU

29. Projected costs determine when biosystems data overdetermination is feasible. In 1984, pre-HGP (fX, pBR322, etc.) 0.1bp/$, would have been $30B per human genome. In 2002, (de novo full vs. resequencing ) ABI/Perlegen/Lynx: $300M vs. $3M 103 bp/$(4 log improvement) Other data I/O (e.g. video) 1013 bits/$

30. Polymerase colonyfluorescent in situ sequencing (FISSEQ) Single pixel sequences Mitra & Shendure

31. Inorganic nanofabrication Diatoms, metalloenzymes, Photoreactions, etc. http://www.urbanrivers.org/diatoms.html http://norrisgroup.uchicago.edu/research/xrayposter/prc.jpg

32. David Goodsell

33. Synthetic Mini-genomes • Digital input > complex atomically precise output • 90kbp genome? All 3D structures known. • 100X faster replication (10 sec doubling) • & selection to evolve widgets & systems? • Utility of mirror-image & other unnatural • polymers. • Chassis & power supply

34. A 90 kbp mini-genome

35. Future of Human Nature Systems Biology & Modeling Speculative Specifications Potential timelines Limits to Mass Limits to Cost (energy & environment) How? Potential pathways to nanotechnology Human biology keeping up: stem cells

38. Faster than exponential

39. The Sims: Sim City’87 Sim Earth’88 Sim Ant’91 Sim Life’92 Sim Farm’93 Sim Isle’95 … Simulation software list

40. Modeling bio-effects on global warming The equatorial Pacific, sub arctic pacific and Southern Ocean are high-nutrient low-chlorophyll (HNLC) areas which may support higher plant biomass if micro-nutrients such as Fe were added... No ocean fertilization study has been long lived enough to follow the effects of iron fertilization through the food web, and hence determine the potential for long term carbon sequestration. models <1% of photosynthetic biomass, phytoplankton ~50% carbon fixation. Chisholm et al. (2001) Science 294(5541):309-1. Dis-crediting ocean fertilization.

41. HIV-1 resequencing New nucleotide sequences processed in GenBank per month (above) Today's total is:

42. D. Kirschner & G Webb Resistance, Remission, & Qualitative Differences in HIV Chemotherapy (Pub) Therapy T(0),Vs(0) in mm-3 = 306, 21 (5.8 years), 217, 31/mm3 (7.7 years), 100, 69/mm3 (8.4 years), 43, 156/mm3 (8.6 years). The rates of exponential increase in Fig. 2a (.03, .02, .01, .005) are inversely correlated to starting CD4+ T-cell counts; decay rates in Fig 2b (all -.2) are not correlated to starting viral levels (different viral set-points would give different values for the parallel slopes) (1,2). The lack of correlation of viral decay rates is an indication of slower clearance of wild-type virus in the external lymphoid compartment. The time to the downward spike in Figure 2b is correlated to starting viral levels (1). The treatment parameters c1=2.0, c2=.17, c3=.15 and the resistance mutation parameter q=10-6 are the same in all four simulations. HIV-AIDS models HIV economic modeling

43. HIV treatement model parameters

44. Vaccines for the 21st Century http://books.nap.edu/html/vacc21/ Level I Most favorable: saves money & Quality-Adjusted Life Years(QALY) Level II < $10,000 < Level III < $100K per QALY saved < Level IV Level I candidate vaccines: Viral: CMV vaccine for 12 year olds, Flu vaccine for 20% of the US per year. Therapeutic vaccines: IDDM diabetes, MS, Rheumatoid arthritis • Bacterial: Streptococcus B & pneumoniae vaccine for infants & 65 year olds. [HIV vaccines prominent already within NIH.] “A quantitative model that could be used by decision makers to prioritize the development of vaccines against a number of disparate diseases” 1985 & 1999.

45. Role of genomics & computational biology in vaccine R&D? DNA vaccines , Intracellular vaccines, RNAi, multiplexed… Gaschen et al. (2002) Science 296(5577):2354-60 Diversity considerations in HIV-1 vaccine selection. Malaria & Mosquito genomes

46. Human Genome Project Ethical, Legal & Social Issues (ELSI) Fairness - Genetic non-discrimination Privacy Reproductive rights - cloning Psychological stigmatization Clinical quality-control Safety and environmental issues - GMO & biowarfare Uncertainties - testing minors Conceptual & philosophical implications - diversity? Commercialization of products - Who owns?

47. Human Genome Project Ethical, Legal & Social Issues (ELSI) Clinton, June 26, 2000. The Genetic Nondiscrimination in Health Insurance and Employment Act of 1999, introduced by Senator Daschle and Congresswoman Slaughter, that will extend these employment protections to the private sector and finish the job of helping to extend protections to individuals purchasing health insurance, begun with the Health Insurance Portability and Accountability Act.

48. ELSI: Do Races Differ? Not Really, DNA Shows Hb variants "evolved to help the ancestors of these groups resist malaria infection, but both prove lethal when inherited in a double dose. As with differences in skin pigmentation, the pressure of the environment to develop a group-wide trait was powerful, and the means to do so simple and straightforward, through the alteration of a single gene. A founder effect explains the high incidence of Huntington's neurodegenerative disease in the Lake Maracaibo region of Venezuela, and of Tay-Sachs disease among Ashkenazi Jews. But Dr. Naggert emphasized that medical geneticists had a much better chance of unearthing these founder effects by scrutinizing small, isolated and well-defined populations, like the northern Finns, the Basques of Spain, or the Amish of Pennsylvania, than they did by going after "races."

49. Dangers of model-free science Lysenko: inheritance of acquired characteristics. "His habit was to report only successes. His results were based on extremely small samples, inaccurate records, and the almost total absence of control groups. An early mistake in calculation, which caused comment among other specialists, made him extremely negative toward the use of mathematics in science. "

50. Dangers of ethics-free science The 1979 release of Anthrax-836 spores in Sverdlovsk. "In 1953 a leak…In 1956, Sizov found that one of the rodents captured in the Kirov sewers had developed a new strain more virulent than the original. The army immediately ordered him to cultivate the new strain…to install in the SS-18s targeted on western cities." Alibek & Handelman "Biohazard" 1999 (Davis) How can we improve our genome engineering tools preferentially toward defense and away from terrorism?