Scientists Create Synthetic Life

Last Updated on Saturday, 17 July 2010 14:57 Written by Mehmet Ali ANIL Friday, 21 May 2010 00:02

I just had to write about this.

Today, a large group in J. Craig Venter Institute, US, seems to engineer their own DNA (or RNA, NA in which there is encoded information, eventually), synthesize it in laboratory, inject it into a bacteria cell. The encoded DNA is structured  so that the bacteria was able to reproduce itself, and the resultant bacterium do also have the reproductive ability, which are two of the key requirements to be a living organism.

Well, this might be one of the days that will go into the textbooks. Ever since I read "The Selfish Gene" by Richard Dawkins, I found the section about "The Replicators" totally mind boggling. The fact that it was self repetition or self organization that is key player of the whole process, a simple algorithm that probably was once in favor of minimizing the energy, or one of millions of such root algorithms, that turned out to result in such large dynamics, as we right now are looking through.

The idea that those key algorithms are being unfolded, piece by piece, that humans are deciphering the motivation, the great kinesis we are right now observing in awe, is enough to give chills. I feel that we're starting to understand organisms in a bottom to up approach, just like mastering the replicator in order to figure genes out, despite of the fact that this was not the approach in biology, and it is not perfectly still.

Right now, that seems to me, genetics is in a point that it will go into a change that should be emphasized, like the turning point that Electronics was once at on 60's, the time the saturation in the need and the theoretical background, resulted in a burst in engineering, an uncontrollable momentum, replication, integration and scaling, such that still even a minuscule amount of enhancement is considered to be an instant realization, even there is a high price to be paid. It didn't stop there, because electronics had this unique property of paving its own way, in other words, it created its own market and design tools by its research and development. Molecular biology seems to be climbing to this peak, once approached its problems with a chemists or a biologist's intuition, and now also with the help of statistical physics and computational advances, is armed with numerical methods and simulations. This saturation of know-how until recetly was encapsulated into the laboratory and reached the public comparably indirectly and under a watchful eye, with generally ethical concerns. But seeing that this breakthrough was achieved by a private institute, if the time of genetic engineering comes with opportunities of easy implementation and integration, the dam is going to fall down. My guess is that we'll see days that the inevitable momentum of engineerable organisms flooding every place the market is in need of, will be overwhelming to every single limiting factor known effective today, religion, ethics, human rights or environmental concerns.

This might seem to be an over-dramatic entry, like those PopSci news, that every small development is presented as great scientific breakthrough, those that make one wonder how come every single day there happens to be one, In fact one should know that I am not presenting this as a huge advance, but It is, in my opinion, a signal for the point that is reached in genetic manipulation, that it is not genetic imitation anymore, it is evolving into a process of encoding, and design, which surely will build its layers of production. There will be ones that will code, ones that will generate blocks of single tasks, ones that will try multitasking, ones that will try to create self-developing and adaptive structures and so on. One should remember that the golden age of Electronics was not the time that the bipolar juction transistor was realized in the lab, it was in contrast, the time that CMOS was perfectly implemented, because it gave the opportunity to divide the workload, segmentize it in a way that one could focus on different parts, the platform that gave the chance for one group to code, one group to write algorithms that the ones that code will use without getting into details of, one group that design the architecture, and others that get into the semiconductor physics. This segmentation did not only gave a chance to divide the labor, but also, speeded up the scaling process, because requirements of every single subunit was evident, which resulted in an environment that the asignee of the problem, remained focused into his expertise.

If molecular biology present similiar circumstances, it will enjoy the benefits of plug'n play, whereas I am still not sure whether we (well, even if we don't want we'll be users of it, it does not seem to be preferential) will enjoy, or will suffer the big step sideways from Darwinian track, exchanging a huge heuristic algorithm with no memory but big sample size, with a combination of scientific methodology and engineering voraciousness.

No matter what, here's something that I will mention years after, when the day will come.

For the ScienceExpress article: http://www.sciencemag.org/cgi/rapidpdf/science.1190719v1.pdf
For the BBC News Report: http://news.bbc.co.uk/2/hi/science/nature/8695862.stm

My Personal LaTeX Template

Last Updated on Saturday, 17 July 2010 14:56 Written by Mehmet Ali ANIL Monday, 17 May 2010 12:34

Since I use LaTeX in order to prepare my documants for a while, I want to share those who are newcomers into the field, the cluster of code that grew sporadically for couple of years, because of my bad copy - paste habit while working with documents. I will at some time create my own functions, but I don't think that would be anytime soon, since I have other works to do. 

Before I begin bombarding you with codes, I use Kile in order to write and compile my code. It is sufficent enough and very user-friendly with its shortcuts and symbol lists.  This is only application I could't part within the whole KDE application list. I use Zotero and Mendeley in order to manage my souorces, and I extract BibTex citations from there. 

First of all, I have a bunch of packages I load everytime I start a document, regardless of the fact that If I use it or not. This is indeed bad habnit, since, packages tend to create conflicts, but to be frankly, who's going to memorize which package is used for what reason? So this is my header:

 
 \documentclass[a4paper,10pt]{article}            % Let's use International standards.
\usepackage{amsmath}            %For math typesetting
\usepackage{amssymb}            %For all those symbols that are generally used in math mode, like logic notation.
\usepackage[utf8]{inputenc}            %In order to map turkish characters written in the code to the turkish characters
\usepackage[turkish]{babel}            %In order to print turkish caharcters, hyphenation, and key words like "Bibliography"
\usepackage{mathrsfs}            %Handwritten fancy character in order to use for e.g. H of Hamiltonian operator.
\usepackage{graphicx}            %Used In order to add figures
\usepackage{lscape}            %Landscape mode
\usepackage{xkeyval}            %In order to create macros
\usepackage{lmodern}            %I use it for its oldstyle numbers
\usepackage[T1]{fontenc}            %For lmodern
\usepackage{subfigure}            %Enables subfigures
\usepackage{gensymb}            %Symbols like Celcius , perthousand etc.
\usepackage{bbm}            %For Complex and Real Set symbols
\usepackage{float}            %In order to bypass figure positioning latex uses, I use [H] for all times
\usepackage{bpchem}            %Chemical Symbols
\usepackage{natbib}            %Bibliography Manager
\usepackage{rotating}            %In order to rotate figures

 Those are pretty self explanatory. I know some may clash, or the ordering might be wrong, but It works anyway. 

Then I define two functions:

 
\let\oldhat\hat
\renewcommand{\vec}[1]{\mathbf{#1}}
\renewcommand{\hat}[1]{\oldhat{\mathbf{#1}}}
\newcommand{\citemali}[2]{\footnote{\citeauthor*{#1}, \citeyear{#1}, #2}} 

The first 2 lines is there in order to change the definition that  is set for \vec which in reality produces: 

in such a way that when \vec is used, it emboldens the symbol used. This bold vector notation can be seen in many books, it is preferred by many modern textbooks.

 The other definition, \citemali is a citation style I saw in Reed College's latex standards page, It creates a superscript number, and adds a footnote to the page, printing a non parenthetical citation.

Until the bug is resolved, the ones that load Turkish Babel must add \shorthandoff{=} after \begin{document} since there is a conflict between graphicx, and babel, I suppose.

 For formulae, I use align, which gives me more freedom in order to present mathematical relations. The & symbol lets you how these formulas are to be aligned. Latex compiler will try to align all those invisible & characters in a line. 

 
\begin{align*}
C\dot V =& I - g_{K}n^4(V-E_k)- g_{Na}m^3(V-E_{Na})- g_{L}(V-E_L)\\
\dot n =& \alpha_n(V)(1-n)-\beta_n(V)n\\
\dot m =& \alpha_m(V)(1-m)-\beta_m(V)m\\
\dot h =& \alpha_h(V)(1-h)-\beta_h(V)h\\
&\alpha_n(V)=0.01\frac{10-V}{exp(\frac{10-V}{10})-1}\\
&\alpha_m(V)=0.1\frac{25-V}{exp(\frac{25-V}{10})-1}\\
&\alpha_h(V)=0.07exp(\frac{-V}{20})
\end{align*} 

 In order to present a figure, I use this codelet, that generates the figure scaled, rotated if necessary, captiones, centered, and positioned rigth in the place it is in the code. 

 
\begin{figure} [H]
\centering
\includegraphics[height=0.9\linewidth,angle=-90]{nullclines.ps}
\caption{Hodgkin Huxley sisteminin I=-10,-30,-60 ve -100 $\mu A/cm^2$ değerleri için çözümler ve her çözüme ait sıfır eğrileri}
\end{figure} 

In order to generate a bibliography, I use these commands:

 
\bibliographystyle{plainnat}
\bibliography{main}
 

 In order to generate a table, I use this code:

 
 \begin{tabular}{c c c c c c }
$I_{dc}$ &$P_{dc}$ & $V_{out}$ & $P_{y}$ &$\eta$ & Distorsiyon\\
\hline\\
74.48&1147mW&4.86V&590mW& 52 & \%48
\end{tabular}
 

Though it is not that latex-worthy in looks, but I have still not found a better one.

 And also, I use the savetrees package if I am sure that there will be no constructive criticism, which shrinks the margins.  In other words, I also add \usepackage{savetrees} to the header.

The last, and probably the least in importance, for reports, I use this scaled titlepage rather than the minimal titlepage latex generates:

 
 \begin{center}
  \vspace{10cm}
  \textsc{\huge Name of the University \\ \bigskip \huge Name of the Faculty}
  \vfill 
  \Huge The name of the course\\ \smallskip
  \Large The topic of the report
  \vfill
  \Large Name, Surname \\ \oldstylenums{Number} \\ \smallskip
  Deney Tarihi : \oldstylenums{The date the lab was attended}\\ Teslim Tarihi : \oldstylenums{\today}
  \vfill
  \Large Grup:\oldstylenums{C09} \\ The research assistant that was in charge: The name of the Research Assistant \smallskip 
\end{center}
\newpage
  

  In my opinion, it is worthy to learn about latex, at least try it. Though you lose time at first, you get to find it handy ans much more beautiful, after getting hang of it. It is a better choice for large documents, like theses, since it gives you much freedom about your style and it does not mess up if a correction is made to the document in the middle afterwards.  

  Good luck to those who get started with LaTeX, and don't hesitate to comment if you would like to reply to this text. 

Jonsi - Go

Last Updated on Saturday, 17 July 2010 14:56 Written by Mehmet Ali Anıl Saturday, 24 April 2010 21:55

This entry is in Turkish only.

Logistic Map in Python

Last Updated on Saturday, 17 July 2010 14:54 Written by Mehmet Ali ANIL Saturday, 24 April 2010 19:52

Whenever I have an assignment to do, I try to do it in a procedure, or a style that I am not familiar with and also that I am eager to learn. This was once typesetting with LaTeX2e, now I am messing around with Python, so instead of realizing Logistic Map in Matlab, I tried to accomplish the task that was given in Python, though I had to put more effort in doing so. 

For the more technically inclined, I am using Python 2.6.4, with Stani's Python Editor, on Ubuntu.  I don't think that would make a lot difference, but for the sake of precision, I wanted to state it. In my opinion, those are a good choice to start, and also I will try Eclipse with Pydev,  since it is an IDE I am more accustomed with.

The Logistic map is a  second order mapping, that is also called a logistics equation in which I will be analyzing as a discrete mapping henceforth. So in our case, 

And we are going to use our logistic map such that the output generated by an nth of a series, will be introduced to another logistic function, so (n+1)th element of the series is obtained. In a more compact way:

that relation causes a simple bevahior when x values evolve with time if r<3. Simply, x values converge into a finite, real value when this relation holds. If r=3, the system undergoes a bifurcation and two different equillibrium points are generated. The behaviour of the system gets more and more complex with r getting larger. At some point, the system behaves chaotically, in other words, the values diverge form the values when x is introduced with an infinitesimally small perturbation. 

I coded in a function oriented fashion, though Python is a language that is designed for an object oriented coding style. The following can also be witthen in C pretty easily, and the runtime would be much less. Also, those are coded in a non-optimized way, I guess I like using Python a bit like Matlab, It eases brainstorming with coding. 

"""
Chaos: Discrete Chaos Module for Python
    Copyright (C) 2010  Mehmet Ali ANIL


    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.


    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.


    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <http://www.gnu.org/licenses/>.
"""
__author__ = "Mehmet Ali ANIL"
__copyright__ = "Copyright 2010 GNU-GPL 3.0"
__license__ = "GNU-GPL 3.0"
__version__ = "0.1"
__status__ = "Production"




import os
import sys
import math
import matplotlib.pyplot as plt
import numpy as np


def LogisticMap(x,r):
    """
    Logistic Mapper
    Input arguments: 
        x: the x[n] variable
        r: the r bifurcation parameter
    Returns:
        x[n+1]
    """
    return x*r*(1-x)


def CreateLogisticSeries(n,x,r):
    """
    Creates a logistic Series, where x[k+1] is calculated form x[k]
    from x[0] till x[n-1] with the bifurcation parameter r.
    Input Arguments:
        x: the x[n] variable
        r: the r bifurcation parameter
        n: length of the list generated
    Returns:
        [ListInput,ListOutout]
    """
    Input=[0]*n
    Output=[0]*n
    Input[0]=x
    for nDummy in range(n-1):
        Output[nDummy]=LogisticMap(Input[nDummy],r)
        Input[nDummy+1]= Output[nDummy]
    Output[n-1] = LogisticMap(Input[n-1],r)
    return [Input,Output]


def LogisticsMapBifurcationDiagram(rInitial, rFinal, StepSize, xInitialLogistic, nLogistics):
    Last50Iterations=[]
    XValueTracker=[]


    for r in np.arange(rInitial,rFinal,StepSize):
        Data=CreateLogisticSeries(nLogistics,xInitialLogistic,r)
        DataOutput = Data[1]
        Last50Iterations.extend(DataOutput[499:999])
        XValueTracker.extend([r]*500)


    plt.clf()
    plt.scatter(XValueTracker,Last50Iterations,marker='o',s=0.005)
    plt.grid(True)
    plt.savefig('/home/mali/Desktop/BifurcationDiagram.png',format='png')
    plt.show()


LogisticsMapBifurcationDiagram(2.5,4,0.001,0.4,1000) 

As a result, one can easily obtain this beautiful bifurcation diagram:

A bifurcation diagram is obtained with plotting persistent slotions, or equilibrium points with respect to the bifurcaion parameter, that was r in our case. Despite of seeing that a very simple mapping rule turns out to be so exotic and complex in action, it was a relief to see that logistics map has values of r, that it returns to acting like it has finite  points. Just like an eye of a tornado, from a distinctive chaotic behaviour, with a little change in parameter r, it intoduces a silent, predictable state space. 

The code provided does a similar thing, but in a rather crude way. I will optimize it, and play with it more, add more chaotic maps to it, and also add the ability to draw Cobweb diagrams.

I encourage you all to spend some time with Python, it is an easy programming language and it has lot of modules  that can be implemented with an easy line of code. Python can be used in an interactive mode, just like Matlab, so that you can play with your data, or alter it in real time. It is not a Matlab alternative in may ways, but it is fun to use it.