1
All I Need is Time! The Modern Theory of
Evolution"Time is the hero of the plot Given
so much time the 'impossible' becomes possible,
the possible probable and the probable virtually
certain. One has only to wait time itself
performs miracles." - George Wald
George Wald (1967 Nobel Prize winner in
Medicine), "The Origin of Life," Scientific
American, vol. 191 1954, p. 46 reprinted on p.
307-320, A Treasury of Science, Fourth Revised
Edition, Harlow Shapley et al., eds., Harper and
Brothers Publishers, 1958. p 309.

• Sean D. Pitman M.D.
• August 2003

2
Evolution

• Definition Common descent with modification
• Mechanism Random mutation and natural selection

3

• Random Mutations
• Are real
• Affect DNA
• Must be inheritable
• Natural selection
• A real force of nature
• A mindless process (no goal in mind)
• Is capable of detecting phenotype changes
• The phenotypic is the functional expression of
  the genotype
• Cannot detect all genotypic changes
• Many genotypic changes are neutral

4
DNA

• Deoxyribonucleic acid
• A chain made up of 4 chemical letters
• Adenine, Thymine, Guanine, Cytosine
• Information is coded by the sequential order of
  these chemical letters
• Similar to other code or language systems
• Morse Code, English, computer code and other
  language systems

5
Transcription and Translation

• Information in DNA is copied into a working
  copy called messenger RNA
• The message is then translated into another
  language system known as proteins
• DNA controls what, when, where and how much
  proteins are produced

6
Decoding DNA

• Three DNA letters (a codon) translates into one
  type of amino acid letter in a protein chain
• Twenty different letters in protein language
• Sequential order of protein letters vital to
  function as with any other language system
• Protein chains are folded into specific
  functional 3D shapes often with the aid of
  other pre-formed proteins

Transcription then Translation
7
Proteins

• Functional units of the cell
• Very specific sequential order and 3D structures
  that are dependent upon the proper sequential
  order
• Some flexibility depending on location and class
  of amino acid letter change
• Changes in DNA and/or the resulting protein
  sequence may be functionally beneficial, harmful,
  or neutral

8
Changes that Arent
9
Gregor Johann Mendel - born July 22, 1822

• Father of genetics
• Studied pea plant hybrids
• Discovered basic principles of genetic
  recombination
• Darwin not aware of Mendels work

10
Genetic Recombination
11
Chromosome Loci

• Gene is a functional stretch of DNA that often
  codes for a protein
• Genes occupy specific positions on chromosomes
  called loci
• For a given locus, different versions of the same
  gene (alleles) will fit or work in this position
  just like different types of wheels will work
  at the wheel location on a car
• If genes arent in the right locus, they will not
  work right (like putting seats where the tires
  are supposed to go)

12
Two Loci Per Trait

• Chromosomes come in pairs
• One chromosome comes from the mother and one from
  the father
• Each locus on one chromosome has a matching locus
  on its sister chromosome
• During genetic recombination, these chromosomes
  must match up exactly and interchange DNA only
  between matching loci

13
Different Pea Traits and Potential Form
Variations

• 1.        Flower color purple or white
• 2.        Flower position axial or terminal
• 3.        Stem length long or short
• 4.        Seed shape round or wrinkled
• 5.        Seen color yellow or green
• 6.        Pod shape inflated or constricted
• 7.        Pod color yellow or green

14
Dominance and Recessiveness

• If maternal and paternal chromosomes have
  different forms at the same loci on the twin
  chromosomes, which form will be expressed?
• Answer The dominant form
• Sometimes there is co-dominance, incomplete
  dominance, or polygenetic governance of a
  particular trait

15
Mating a purely Dominant with a purely Recessive
line
Phenotype
Genotype
Homozygotes and Heterozygotes
16
The Static Gene Pool

• Fantastic potential for expressed variations
• Made up of all the genes in a population
• Many different options (alleles) per gene locus
  exist in a gene pool
• Genetic recombination produces random allelic
  swapping at specific loci, but it does not
  produce new alleles or new loci
• i.e., a red bike with knobby wheels or a yellow
  bike with thin wheels or a red bike with wide
  wheels etc., but the basic bike setup would not
  change nor would the variation potential change
  via part recombination

17
The Evolving Gene Pool

• Must go beyond what genetic recombination can
  offer
• Dependent upon random mutation to existing DNA
  sequences to create new alleles and genetic loci
  in the gene pool of options

18
The Potential Space Checkerboard of Sequence
OptionsA Comparison to the English Language

• On our checkerboard each square represents a
  different potential sequence of a given length
• For example, the total number of potential
  3-letter sequences is 17,576. Each one of these
  sequences is represented by a different square on
  our checkerboard.
• Some squares represent sequences that are
  functionally beneficial, others are functionally
  detrimental, but most are functionally neutral

19
Evolving . . . Small Words

• 3-letter checkerboard size 17,576 squares
• 2-letter checkerboard size 677 squares
• Defined 3-letter words 972 (ratio 118)
• Defined 2-letter words 96 (ratio 17)
• A random change in a 2- or 3-letter sequence will
  often produce a new meaningful word
• cat hat bat bad bid did dig
  dog

http//www.yak.net/kablooey/scrabble.html
http//www.aivazoglou.gr/2.html
20
Evolving . . . Bigger Words

• 7-letter checkerboard size 8,031,810,176 squares
• Meaningful 7-letter words 23,109
• Meaningless 7-letter sequences 8 Billion
• Each meaningful word is surrounded by 347,561
  meaningless sequences like an ocean of squares
  on our checkerboard
• The likelihood that a random change to a given
  7-letter sequence will end up on a meaningful
  word is not nearly as good as it was for 2- and
  3-letter sequences
• Much more time is required to find new 7-letter
  words on our checkerboard of sequence options

http//www.aivazoglou.gr/7.html
21
Moving Up to Short Phrases

• Methinks it is like a weasel Shakespeare
• This sentence is 28 characters in length
• Checkerboard size 10e40 squares (27e28)
• If each square were 2cm across this checkerboard
  would be 1,000 trillion km across (621 trillion
  miles)
• 93 million miles to the sun
• 25 trillion miles to nearest star (Alpha
  Centauri)
• How many meaningful 28-character phrases are
  there in the English language?
• Im not sure perhaps a trillion?
• If there were a trillion, each meaningful phrase
  would be surrounded by a vast ocean of ten
  billion trillion trillion meaningless phrases
• Changing one letter per second, with no repeats,
  it would take over 300 million trillion years to
  come across just one of the trillion meaningful
  phrases on our checkerboard
• Math 10e40 / 10e12 / 60 / 60 / 24/ 365 3.79 x
  10e20

22
Increasing the PopulationIt does help but how
much?

• Take the same scenario as before and increase the
  number of evolving phrases a trillion fold
• This would search out the squares on our
  checkerboard a trillion times faster
• Still works out to be over 300 million years to
  find just one of one trillion meaningful phrases
• Note We arent even talking about beneficial
• phrases here - just meaningful
  phrases

23
Expanding Checkerboardvs.Expanding Population

• If the average distance between any two
  meaningful sequences is two character changes
  (two steps on the checkerboard), the average
  random walk distance for one individual is more
  than 729 steps (27e2).
• Note also that random genotypic walk cannot be
  guided by a phenotypic selection process
• As one moves from one meaningless square on the
  checkerboard to another meaningless square, no
  change in meaning takes place, and therefore no
  change in phenotype - such changes are referred
  to as neutral mutations

24

• For the same 2-mutation gap
• If the population were increased to 1,000
  sequences, the random walk for this population as
  a whole would average less than 1 step
• Again, increasing the population decreases the
  average time required for the evolution of new
  meaningful sequence codes
• But, can the population expansion keep up as the
  average gaps between meaningful sequences expands?

25
Doubling the Gap

• Lets expand the average gap from 2 to 4 character
  changes or steps on our checkerboard
• Now, instead of 729, there are 531,441
  meaningless squares for every one meaningful
  square
• With each doubling of the neutral gap, the
  potential space increases by a factor of 2
• In order to evolve at the same rate, our
  population would also have to increase in size by
  a factor of 2

26
Doubling a Few More Times

• A gap of 8 282 billion meaningless squares on
  the checkerboard
• A gap of 16 79 billion trillion . . .
• A gap of 32 1 billion trillion trillion
  trillion . . .

Note The junk-squares on the checkerboard
quickly outpace the ability of a given
population to keep up with the
expanding neutral gaps found at higher
and higher levels of functional complexity.
27
Comparison to Real Life?

• A gap of 32 Amino Acids 4.29 x 10e41 (100
  thousand trillion trillion trillion)
• Total bacteria on Earth 5 x 10e30 (5 million
  trillion trillion)
• A checkerboard with 10e41 meaningless AA squares
  divided among 5 million trillion trillion
  bacteria would require each individual bacterium
  and its offspring (just one in a steady state
  population) to undergo a random walk of around 85
  billion steps before success would be realized
• Time per step 10 years
• Based on a very high mutation rate of 10e-5 per
  sequence per generation (one mutation every
  100,000 generations) with a generation time of 1
  hour
• Average time to success 850 billion years

http//news.bbc.co.uk/1/hi/sci/tech/158203.stm
28
Real Life Examples of Gene Pool Evolution?

• Antibiotic Resistance - yes
• Single Protein Enzymes - yes
• Multiple Protein Systems - not yet
• Alife Computer Software - sorta

29
De Novo Antibiotic Resistance

• Is a real example of evolution in action
• Is the result of mutation and natural selection
• Is based on antibiotic-target specificity
• Streptomycin 30S ribosomal subunit
• Erythromycin 50S ribosomal subunit
• Fluoroquinolones DNA gyrases and topoisomerases
• Rifampin beta-subunit of RNA polymerase
• Penicillins penicillin binding proteins (cell
  wall synthesis)
• Vancomycin stem peptides involved with
  cross-linking
• Isoniazid inh enzyme involved with mycolic acid
  synthesis

30
Antibiotic-Target Interference

• High ratio of interfering sequences in sequence
  space (like ratio in 2 and 3-letter words in
  sequence space)
• Much easier to interfere with or break a previous
  interaction/function than to create a brand new
  function
• Remember Humpty Dumpty and all the kings men!

31
Antibiotic Evolution is Predictably Rapid

• The higher the antibiotic-target specificity, the
  easier it is to interrupt this interaction
• Odds are good that a random mutation will result
  in some degree of interference rapid antibiotic
  residence
• In real life, this is exactly what happens
• Antibiotic resistance evolves extremely rapidly
  in all types of bacteria when exposed to an
  antibiotic
• The same principle holds true for mechanisms of
  drug resistance in other life forms (malaria,
  rats, weeds, etc.)

32
Up a Level to Single Protein Enzymes

• Have independent function not dependent upon the
  disruption of a pre-established
  function/interaction
• Specific AA sequence and 3D orientation vital to
  enzyme function
• If sequence changed too much, all function will
  be lost

33
Invariant Amino Acids

• Cytochrome C (100 AA)
• Comparisons of 40 species 35 invariant AA
• Humans and Bread Mold - 66 Identical
• 40 more are restricted between 2 or 3 AA options
• similar in chemical character hydrophilic,
  hydrophobic, basic acidic, or neutral with
  respect to water
• Only a few positions can tolerate a wide range of
  AA substitutions

34
100aa Checkerboard

• 10e130 squares on our 100aa checkerboard
• One trillion trillion trillion trillion trillion
  miles wide
• Note that there are only 10e80 atoms in the
  visible universe!
• How many squares with cytochrome c function?
• High estimates 10e60 (lower estimates 10e40)
• Cytochrome C vs. Total At least 1 in 10e70
• How many other functions requiring at least 100aa
  would be beneficial to a given life form in a
  given environment?
• Human genome thought to have around 35,000
  protein coding genes
• 35,000 times 10e60 is only 10e64
• 10e64 nonfunctional sequences for every
  functional sequence

http//www.nso.edu/sunspot/pr/answerbook/universe
.html
35
But Some Enzymes Have Evolved in Real Time!

• Barry Hall and Lactase Evolution in E. coli
• LacZ genes deleted in E. coli
• Code for tetramer lactase enzyme of 1,000aa per
  subunit
• Mutant bacteria grown on lactose enriched media
• In one or two generations, the lactase function
  evolved
• Completely different gene (ebg gene hexamer of
  1000aa per subunit) evolved the lactase function
  with a single point mutation
• What are the odds?!

36
Double Mutant E. coli

• Hall deleted both the LacZ and ebg genes from
  certain colonies of E. coli
• Despite being grown on lactose enriched media for
  tens of thousands of generations, while
  undergoing high mutations rates, the lactase
  function was never evolved
• Hall described these bacteria as having, limited
  evolutionary potential
• Note These same bacteria would quickly evolve
  resistance to any antibiotic environment

37
Limited Evolutionary Potential

• Many species of bacteria cannot evolve the
  lactase function despite hundreds of thousands of
  generations of records
• Salmonella, Proteus, Pseudomonas etc.
• Lactase function would be beneficial if it
  evolved
• What is the limiting factor?
• A low ratio of beneficial functions vs. potential
  functions at this level of complexity (1,000aa
  level with around 10e1300 options)?

38
It Only Gets Worse

• There are quite a few examples of single protein
  functions evolving (nylonase, lactase, antibody
  enzymes, enzyme cascades etc.)
• However, there are NO examples of multi-protein
  functions evolving - period
• Where multiple proteins are required to work
  together at the same time in a specific
  orientation with each other (i.e., bacterial
  motility systems such as the flagellar apparatus)

39
The Famous Flagellum

• 50 or so different proteins working in concert
• Each protein part averages a few hundred amino
  acids
• Total sequence length Probably more than
  20,000aa
• Sequence space At least10e26,000 potential
  differences
• So, what is the ratio of all beneficial sequences
  requiring at least 20,000aa compared with the
  total of 10e26,000 potential sequences at this
  level of complexity?

40
Flagellar Subsystems

• Parts of the flagellum used in other systems of
  function
• Similar proteins in type III secretory systems
  (TTSS)
• TTSS coded for by ten genes each homologous to
  genes in the bacterial flagellum code (The
  flagellum requires an additional 30 or 40 unique
  genes)
• Nguyen et al. 2000 suggest that TTSS evolved from
  the flagellum and not vice versa
• TTSS only found in animal and plant pathogens
• Argument is that TTSS could only have existed
  since the evolution of animal and plants so it
  evolved from the flagellum
• TTSS genes can be carried on plasmids and
  transferred horizontally flagellar genes are
  not carried on plasmids

41
ALL the Right Parts

• What if all the needed parts to make a flagellum
  already existed as parts of other systems of
  function?
• Whenever a flagellum was needed, these parts
  would just self-assemble to make a flagellum
  since all the parts are there right?
• Would a watch self-assemble if I put all the
  necessary watch parts in a bag and shook the bag
  for a billion years?

42
Part Assembly

• Remember, the information in DNA controls not
  only part production, but the timing and location
  of part production.
• DNA is like a biochemist who synthesizes complex
  compounds. In order for the synthesis to be
  successful, the sequential order and amount of
  chemicals added to and removed from solution is
  vital to the formation of the final product.
• The parts, by themselves, without this higher
  informational input, will not self-assemble to
  produce much of anything.
• Example All the necessary amino acids are always
  there to produce any protein that would be
  beneficial to the cell. But, left to themselves,
  the AAs do not self-assemble in a functional way
  that is greater than the sum of their parts.

43
Wishful Speculations

• All of this current hypothesizing about how
  evolution must have happened is highly
  speculative, and accounts for cell biologist
  Franklin Harold's frank admission
• "There are presently no detailed Darwinian
  accounts of the evolution of any biochemical or
  cellular system, only a variety of wishful
  speculations."

Harold, F. 2001. The Way of the Cell Molecules,
Organisms and the Order of Life. New York
Oxford University Press.
- William Dembski
http//www.designinference.com/documents/2003.02.M
iller_Response.htm
44
Computer Evolution

• Lenski et. al., The Evolutionary Origin of
  Complex Features, Nature, May 2003
• Goal - the evolution of logic functions
• Population - 3600 individuals divided among 50
  different genomes
• Each individual had 50 lines of meaningful
  computer code
• 15 lines of code required for the
  self-replication function
• 35 lines were repeated copies of a no operation
  code (no logic function)
• Random translocation mutations of lines and
  portions of lines of code

• The evolution of specific logic functions
  rewarded with increased reproductive fitness
• Generations 15,873
• Result 23 of 50 genomes evolved the highest
  logic function (EQU)

45
A Few More Details

• The size of the checkerboard 5.6 x 10e70 number
  of options
• Based on ancestral genome length of 50 with 26
  possible instructions at each site
• Size of gap between starting code and EQU
  function 16 steps/mutations
• Other beneficial intermediate functions NAND,
  AND, OR, NOR, XOR, NOT
• Average gap between all functions 2.5 steps
  (search space 3,400 options between each step
  every option quickly covered with a population of
  3,600 individuals)
• Average number of steps required to achieve EQU
  function in the 23 successful genomes 103
• 45 beneficial, 18 detrimental, 48 neutral

46
Removing Intermediate Steps

• Lenski also experimented with an evolutionary
  scenario where only the EQU function was defined
  as beneficial
• Result NONE of the 50 populations evolved the
  EQU function
• As expected, these populations tested more of the
  potential sequence space than those in the
  reward-all environment (2.15 x 10e7 vs. 1.22 x
  10e7 Plt0.0001, Mann-Witney test)
• Still, this was a minute 1 x 10e-60 of the total
  space on the checkerboard
• This is like searching out a grain of sand when
  the entire universe remains unexplored
• The EQU function could be on any grain of sand
  anywhere in the universe
• Without the intermediate functions arbitrarily
  defined as beneficial, the resulting neutral gap
  of 16 steps was simply impossible for Lenskis
  population of computer creatures to get across
• Even in trillions upon trillions upon trillions
  of years.

47
All Evolution Needs is TimeYeah!A WHOLE
LOTTA TIME!!!