#+TITLE:Adaptive Reading Group
#+OPTIONS: toc:2 num:nil ^:nil
#+EXPORT_EXCLUDE_TAGS: hide
* Reading
** April 9th
- jthsmithGA.pdf
- decisionmaking.pdf
** Axiomatization of Computability, Churches Thesis
:PROPERTIES:
:CUSTOM_ID: axiom-of-comp
:END:
axiomatization-of-computability-churches-thesis.pdf
- Church's Thesis
- wiki:Lambda_calculus
- effectively calculable :: function of positive integers is such if
it can be identified with the notion of a recursive function of
positive integers (or of a lambda-definable functions of positive
integers)
some (like Post) thought that Church's /definition/ of effectively
calculable placed limits on the mathematical processes available to
humans and was too quick in defining such a process when we should be
continually trying to re-evaluate and expand the process.
|
\-> is a human nothing more than a Turing machine in terms of
complexity? Stephan Wolfram would argue yes.
- partial function :: a function which is not defined for all inputs
(i.e. division)
- Kolmogorov Machine :: an alternative to a turing machine with a type
of connected graph for storage
** Does Efficiency Sensing Unify Diffusion and Quorum Sensing
:PROPERTIES:
:CUSTOM_ID: does-efficiency-sensing-unify-diffusion-and-quorum-sensing
:END:
Does Efficiency Sensing Unify Diffusion and Quorum Sensing
straightforward and largely uninteresting simplification of existing theories
efficiency sensing measures the combination of (cell density,
diffusion properties, and spatial distribution) and it is the
combination of these three (basically how much can I saturate my
surroundings) that really matters
*** intro
proposes /efficiency sensing/ as a unification of /diffusion sensing/
and /quorum sensing/. avoids problems of honesty and complex
environments through growth of cells in microcolonies.
certain bacterial behaviors only happen on make sense at high cellular
densities. the cells seem to signal/sense each other to make these
determinations.
- bacteria secrete signaling molecules which disperse in the
environment
- often these chemicals are /autoinducers/
- all responses to high levels of autoinducers tend to only make sense
when carried out by many bacteria
- quorum sensing /QS/ :: the use of autoinducers to both determine if
enough bacteria are present to act, and to coordinate the action
- diffusion sensing /DS/ :: comeback to quorum sensing, basically the
idea is that the only real information returned by autoinducers
is their rate of diffusion (whether caused by non-porous
surroundings, production by a quorum of other bacteria, or
production by third parties)
- DS is simpler than QS because it doesn't rely on assumptions about
social behavior
this paper will
- review QS and DS
- show that /distribution/ is more important than /density/ for
autoinducers.
- propose *efficiency sensing* ES as a combination of QS and DS
- show that microcolonies default some problems of complexity and
cheating
*** problems with QS
QS problems /in situ/
- complexity problems
- spatial heterogeneities :: non-uniform environment
- biodiversity :: issues of cross-talk from other cells
- cheating problems
- mutants that consume but do not produce /public goods/ can
out-compete normal individuals
- one /possible/ answer to this is that when grown in microcolony
(live near your children) the hereditary sharers would share
mainly with their children
*** QS v.s. DS
there are some situations when the autoinducer is clearly used to
sense a quorum (e.g. bio-luminescence where the action take place
internally, but only makes sense with many neighbors)
*** ES hypothesis
cells can only measure the combination of density, distribution, and
autoinducer medium permeability
autoinducer concentration is a better predictor for effector
concentration than it is for either /diffusion/ or /quorum/
as cost saver
- easier to produce autoinducers than effectors
- low autoinducer production needed if its easily detected
- many different effectors may need to be produced
*** microcolonies avoid problems
growing in clumps
- promotes inter-clump diversity
- decreases likelihood of intra-clump cheating
*** future work
more studies /in situ/ rather than with bacterial dissolved in liquids
*** my thoughts
- if the /system/ at large is arbitrary, and takes into account
physical inconsistencies (i.e. the system could be a single cell in
a box) then the QS and DS may be the same, namely a /quorum/ is
simply enough cell in the environment to create the required density
of autoinducer.
- seems that QS and DS are setup as straw men
- its not fair to discriminate between /spatial distribution/ and the
/cell density/ which QS purports to measure. If you draw your
bounding box correctly then spatial distribution inconsistencies no
longer matter
#+begin_src ditaa :file data/distribution-vs-density.png :cmdline -r :cache :exports results
+-----------------------------------------------------------------+
| |
| +------------------------+ |
| | | |
| | | + |
| | + + | |
| | + | |
| | | |
| | + | |
| | + + | |
| | | |
| | yes quorum here cF78 | |
| +------------------------+ + |
| |
| + |
| |
| |
| + |
| |
| |
| no quorum here + c1AB |
+-----------------------------------------------------------------+
#+end_src
#+results[c13aedcf7ddb7427e965070ed7a08d88d9e3cc0b]:
file:data/distribution-vs-density.png
perhaps it would be better to give each cell its own density metric
at some radius