Delila Program: ev

ev program

Documentation for the ev program is below, with links to related programs in the "see also" section.

{  version = 4.13; (* of ev.p 2011 Sep 28}

(* begin module describe.ev *)
   ev: evolution of binding sites

   ev(list: out, all: inout, evp: inout, output: out)


   list: a record of the evolutionary events that occurred.  See the
      description of evp for the kinds of data that can be printed.  The data
      may be graphed as desired using the xyplo program.  The first few lines
      of the file form an informative header.  All of these lines begin with
      an asterisk, '*', which acts as a comment.  The data themselves are
      organized into individual lines broken by spaces into a set of tokens.
      These are described in the header.

   all:  All the variables, genome sequences and genetic structure to allow
      continuation of the evolution.  Start this as an empty file.  Under
      unix you can use:

         touch all

      After the ev program has been run, the all file will be filled with the
      last generation.  If you run ev again, it will use the current all file
      to continue.  To make a movie, step by 1 or more generations and use
      evd to extract information from the all file.  When you want to start
      over, clear the file out.  To do this under unix you can:

          echo -n "" > all

   evp: parameters to control the program, one per line:

      parameterversion: The version number of the program.  This allows the
         user to be warned if an old parameter file is used.
         When possible, the program will upgrade old parameter files.

      Number of creatures (c, integer).

      Number of potential binding sites per creature (G=Genome, integer).

      Number of sites per creature (gamma, integer).

      Width of the recognizer in bases (width, integer).

      Bases per integer of the recognizer weights (bpi, integer).

      Mutation rate.  The original method gives the rate in hits per
      creature per generation (mu, integer).

         New method as of 2006 Jun 28: mutations can be specified in
         two ways: per genome or per base.

         This is determined by the first character on the line.
         if the first character is:

           'g': mutations per genome per generation
           'b': mutations per base   per generation

         If the first character is a digit or a space, then the
         original method (g, mutations per genome) is used.

         Each of these is followed by an integer.

         When mutations per base are given, the form is mutations
         in a defined length of sequence.

         Method: Provide mutations per base as a real number: "1 in
         every ______ bases" Compute: m = genome size * mutations /
         bases. To create mutations in an organism, split m into
         integer (mi) and decimal (md) parts. Do the integer part of m
         mutations. Chose a random number r between 0 and 1. If r <=
         md, do an additional mutation. In this way the requested
         mutations will be done on average.

         Note:  The original method was implemented for speed.
         However, it is not natural in the sense that polymerases and
         other factors work on a per base basis.  So the new method is
         more reasonable to use if one is exploring effects of genome

      Seed: a real number between 0 and 1 used to start the random
         number generator.  The date and time is used if this
         number is outside 0 to 1.

      Cycles: number of additional generations to run (cycles, integer).

      Display interval:  for example, 10 means every 10th generation.

      Display control: the first 7 characters on the line control
         the kind of data printed to the list file:
             a = display average number of mistakes and the standard
                 deviation for the population.
             c = display changes in the number of mistakes.
                 The current Rsequence is given if r (below) is turned on.
                 This allows graphs of Rsequence vs mistakes to be made.
             g = display genomic uncertainty, Hg.  If this deviates much
                 from 2.0, then the model is probably bad.
             i = display individuals' mistakes
             o = display orbits: information of individual sites is shown
             r = display information (Rsequence, bits)
             s = current status (range of mistakes) is printed to the output
             m = current status (range of mistakes) is printed to the list

         These may be in any order.   Any other characters (eg, blanks) are

      Selecting: boolean. If true (ie the first character is 't'), then the
         organisms are sorted by their mistakes.   If false, then the
         organisms are randomly sorted.  Normally this should be 'true', but
         it does allow one to switch the selection off suddenly and watch
         random drift (with no evolution) and the decay of existing patterns
         by entropy increase.  Selecting is true unless the first character
         on the line is an 'f'.

      StorageFrequency:  The frequency (every so many generations) with which
         to store a copy of everything in the all file.  If the computer
         crashes part way through a long run, then the run can be continued
         from the last storage.  Of course, a storage is always made at the
         end of the evolution.

      RandomSites:  If the first character is 'r' then the gamma sites are
         placed randomly.  'n' means that the sites are placed randomly,
         but not overlapping.  'a' means that the sites are placed in a
         regular array with constant spacing.  In either case sites are
         placed outside the recognizer gene, and there are exactly gamma
         sites.  That is, two randomly placed sites cannot be in the same
         place.  This avoids conflicts that would make interpretation of the
         results difficult.

      dummy: real.
         This function has been removed.  Just use 0 as the value.

      specialrule: char.  SPECIAL RULE parameter.  This parameter (introduced
         [2001 June 6]) determines what to do when two organisms tie for
         survival.  If the first character on the line is:
         r - random survival of one organism
         b - both survive (original method)
         s - one survives depending on the sorting algorithm.  This is an
             arbitrary function of how the quicksort routine works and is
             therefore generally not a good way to make the decision.
         See below for more details.

      haltoncondition: char.  This parameter (introduced
         [2006 June 24]) causes ev to halt when a given condition
         has occured.
         If the first character on the line is:
         - none   - no halting condition
         r Rs>=Rf - the best creature has Rs at least equal to Rf
         m mistakes 0 - the best creature makes no mistakes
         b both r and m

      Note: As of version 3.68 the program can automatically upgrade
         the parameter file.

   output: messages to the user, including warnings about conditions,
      If the display control in evp (see above) includes 'o', then
      the generation number and the range of mistakes are given.
      If the display control includes 'a', then the mean and standard
      deviation of the mistakes are also given.


   A population of evolving creatures is modelled.  Each creature
   consists of a genome made of the 4 bases.  All creatures have a
   certain number of binding sites, and the recognizer for the sites
   is encoded by a gene in each genome.  The genomes are completely
   random at first.  The recognizer of each creature is translated
   from the gene form to a perceptron-like weight matrix.  This matrix
   is scanned across the genome.  The number of mistakes made is
   counted.  There are two kinds of mistake:

       how many times the recognizer misses a real site


       how many times a non-site is detected by the recognizer.

   These are weighted equally.  (If they were weighted differently it
   should affect the rate but not the final product of the model.)
   All creatures are ranked by their number of mistakes.  The half of
   the population with the most mistakes dies; the other half
   reproduces to take over the empty positions.  Then all creatures
   are mutated and the process repeats.

   The integer weights of the recognizer are stored as base 4 numbers
   in twos complement notation.  a=00, c=01, g=10, t=11.  If 'bases
   per integer' were 3, then aaa encodes 0, acg is 6, etc.  txx and
   gxx (where x is any base) are negative numbers; ttt is -1.

   The threshold for recognition of a site is encoded in the genome
   just after the individual weights.  It is encoded by one integer.


   SPECIAL RULE: if the bugs have the same number of mistakes, reproduction
   (by replacement) does not take place.  This ensures that the quicksort
   algorithm does not affect who takes over the population, and it also
   preserves the diversity of the population.  [1988 October 26]  Without
   this, the population quickly is taken over and evolution is extremely

   [2001 June 6] In response to William Dembski's objection (see below for a
   link) that this rule is inserting information into the results, a new
   parameter is now available to turn off this rule.

   What does this rule mean?  It means that in a duel it is possible for two
   bugs to have a tie.  This is not an unreasonable result, and it frequently
   happens in the natural world!  What does removing the rule mean?  It means
   that the bugs will duel to the death, even if it is an arbitrary death!
   Clearly this cannot affect the overall evolution, but it might slow things
   down.  Indeed it is interesting because it is often the case that in
   fights the combatants are not killed.


   TO RUN THE PROGRAM, create an empty all file (see the description of the
   all file for how to do this) and obtain a copy of an evp file (see the
   "see also" section below for links to examples).  Note that in parameter
   files everything to the right of the actual parameter is ignored on each
   line.  Do not delete this material, it serves as a comment to remind you
   what the parameter does!  Then you can run the ev program.  The results
   are in the list and all files.  To see the contents of the all file (which
   will be a binary on your computer) use the evd program.  You can use the
   xyplo program to plot items from the list file.

   HISTORICAL NOTE:  After getting the idea that Rsequence is close to
   Rfrequency from data on ribosome binding sites (on 1982 December
   7), it became clear that this had to have evolved.  Eventually I
   recognized that Rsequence must evolve toward a more or less set
   Rfrequency.  (See the paper Schneider.ev2000 for a discussion of
   this.)  The challenge was to test this hypothesis.  My first
   thinking was that a model of the evolution would take thousands of
   years on a computer.  However, I read the book "Ever Since Darwin,
   Reflections in Natural History" by Stephen J.  Gould and found in
   there an interesting chapter "Is the Cambrian Explosion a Sigmoid
   Fraud?" (Gould1977.sigmoid)  This described Gould and Eldredge's
   idea of punctuated equilibrium.  I realized that with selection, if
   an organism has an advantage it can take over a population
   exponentially.  By 1984 I had written up my thesis (on the
   observation that Rs ~ Rf, Schneider1986) and had a week before my
   defense in which I had nothing more to do.  So I wrote the ev
   program and it showed that Rs does indeed evolve toward Rf.  I
   don't recall that I even mentioned this during the defense, but it
   was a nice backup!


   For information calculations, see:

   author = "T. D. Schneider
    and G. D. Stormo
    and L. Gold
    and A. Ehrenfeucht",
   title = "Information content of binding sites on nucleotide sequences",
   journal = "J. Mol. Biol.",
   volume = "188",
   pages = "415-431",
   year = "1986"}

   The inspirational article is:

   author = "S. J. Gould",
   title = " Is the Cambrian Explosion a Sigmoid Fraud?",
   booktitle = "Ever Since Darwin, Reflections in Natural History",
   publisher = "W. W. Norton \& Co.",
   address = "N. Y.",
   name ="Stephen Jay",
   pages = "126-133",
   year = "1977"}

   Applying the Perceptron (a simple neural net) to binding sites:

   author = "G. D. Stormo
    and T. D. Schneider
    and L. Gold
    and A. Ehrenfeucht",
   title = "Use of the {`Perceptron'} algorithm to distinguish translational
   initiation sites in {{\em E. coli}}",
   journal = "Nucl. Acids Res.",
   volume = "10",
   pages = "2997-3011",
   year = "1982"}

   The evolution paper describing this program is:

   author = "T. D. Schneider",
   title = "Evolution of Biological Information",
   journal = "Nucleic Acids Res",
   volume = "28",
   number = "14",
   pages = "2794-2799",
   year = "2000"}


A lovely (but long) evolution can be had with the following evp:

3.42  version of ev that this parameter file is designed for.
40000 CYCLES
cgrs5678 a=av, c=change, i=indivls, g=Hg, r=Rs, o=orbit, s=status, m=mistakes
a     r = random site placement, n = no overlap random, a = array
0     sigma
b     SPECIAL RULE for ties: r: random, b: both, s: sort
-     halt on condition: - none; r Rs>=Rf; m mistakes 0; b both r and m

The list file may be plotted with this parameter file for xyplo, the xyplop:

3 6       zerox zeroy         graph coordinate center
x 0 40000 zx min max          (character, real, real) if zx='x' then set xaxis
y -1 6    zy min max          (character, real, real) if zy='y' then set yaxis
10 28 1 1 xinterval yinterval xsubintervals ysubintervals: axis intervals
8 8       xwidth    ywidth    width of numbers in characters
0 2       xdecimal  ydecimal  number of decimal places
15 15     xsize     ysize     size of axes in cm
Rsequence (bits) | Hg (bits) near 2 | mistakes/gamma are connected circles
a         zc                  'c' crosshairs, axXyYnN
n 2       zxl base            if zxl='l' then make x axis log to the given base
n 2       zyl base            if zyl='l' then make y axis log to the given base
1 3       xcolumn   ycolumn   columns of xyin that determine plot location
2         symbol column       the xyin column to read symbols from
0  0      xscolumn  yscolumn  columns of xyin that determine the symbol size
0 0 0     hue saturation brightness   columns for color manipulation
          symbol to plot      'c'=circle, 'b','d'=box, 'x', '+', 'I', 'f', 'g'
r         symbol flag         character in xyin that indicates that this symbol
0.05      symbol sizex        side in inches on the x axis of the symbol.
0.05      symbol sizey        as for the x axis, get size from yscolumn
cl 0.05   connection (example for connection is c- 0.05 for dashed 0.05 inch)
n  0.05   linetype  size      linetype l.-in and size of dashes or dots
          symbol to plot      'c'=circle, 'b','d'=box, 'x', '+', 'I', 'f', 'g'
g         symbol flag         character in xyin that indicates that this symbol
0.05      symbol sizex        side in inches on the x axis of the symbol.
0.05      symbol sizey        as for the x axis, get size from yscolumn
cl 0.05   connection (example for connection is c- 0.05 for dashed 0.05 inch)
n  0.05   linetype  size      linetype l.-in and size of dashes or dots
c         symbol to plot      'c'=circle, 'b','d'=box, 'x', '+', 'I', 'f', 'g'
c         symbol flag         character in xyin that indicates that this symbol
0.05      symbol sizex        side in inches on the x axis of the symbol.
0.05      symbol sizey        as for the x axis, get size from yscolumn
cl 0.05   connection (example for connection is c- 0.05 for dashed 0.05 inch)
n  0.05   linetype  size      linetype l.-in and size of dashes or dots
          **** version 8.50 of xyplop: uses cm for distances
. 0 6 0.10
. 0 5 0.10
- 0 4 0.10
. 0 3 0.10
. 0 2 0.10
. 0 1 0.10
l 0 0 1

Note that dotted lines are placed across the graph for each bit, but a
dashed line is put for Rfrequence (= 4 bits).  The vertical axis represents
the three kinds of data in the list file.

see also

   Program to display the current population: evd.p
   Program to graph the list file: xyplo.p

   A very early original test run:

       parameter file ev.original.evp
       list           ev.original.list

        output from the evd.p display program:
       display        ev.original.display
       evfeatures     ev.original.evfeatures
       genomes        ev.original.genomes
       sites          ev.original.sites

   Example ev parameter file: ev.example.evp
   Example xyplop parameter file given above: ev.example.xyplop

   Parameter file for selective phase in the paper: evp.selection
   Parameter file for atrophy   phase in the paper: evp.atrophy

   The "Evolution of Biological Information" paper (with active hypertext
   links in references):

   The "Evolution of Biological Information" paper at Nucleic Acids

   image for donor.pure TO MAKE LOGOS:
   Briefly, use the evd.p program to get the sequences (in the sites file)
   followed by makebk.p (preferably in automatic mode) to create a Delila
   book.  Then use the delila system to create the logo:  More details are given in
   the evd.p program documentation.





   William Dembski's objection:

   Page testing William Dembski's claim:


   Thomas Dana Schneider


   none known (well, actually it's full of them... ;-)

technical notes

   RANDOM NUMBERS: The program can use the date and time to create random
   numbers.  These are computer system calls and may not function correctly
   on other computers.  One solution is to simply delete the calls to disable
   them if you are having trouble compiling.  You can always set different
   seeds to start the random numbers by hand.  See the link above for modules
   to change the time function for different compilers.

   SPEED IMPROVEMENT: 2002 April 3.  I noticed that not all cases of passing
   the e: everything were by var.  They were being passed by value.  Since
   this means that a *copy* of "everything" is made for each call, this
   should be very expensive!  I ran the program (version 3.71) on an
   evp.original and it took 9.5 or 9.6 user seconds on my Sun Ultra 60
   according to the Unix time function.  I then changed the variable e:
   everything in the dorseq procedure to be var (passed as a variable
   pointer) and the time dropped to 5.7 seconds!  Here are the successive
   improvements for the changes:

      no change      9.51+/-0.08 seconds in 10 tries
      dorseq         5.72+/-0.04 seconds in 10 tries   !!!
      puteverything  5.74+/-0.05 seconds in 10 tries
      listparams     4.12+/-0.04 seconds in 10 tries
      dori           4.05+/-0.05 seconds in 10 tries
      dogenomich     2.01+/-0.03 seconds in 10 tries   !!!
      makemoreheader 2.01+/-0.03 seconds in 10 tries
      putout         2.00+/-0.02 seconds in 20 tries (2.005 seconds)
      dori           2.01+/-0.03 seconds in 20 tries change fbl

   Wow, nearly a 10 fold speed up ...

(* end module describe.ev version = 2.50; (@ of ev 1988 oct 6 *)
{This manual page was created by makman 1.45}

{created by htmlink 1.62}
U.S. Department of Health and Human Services  |  National Institutes of Health  |  National Cancer Institute  |  | 
Policies  |  Viewing Files  |  Accessibility  |  FOIA