C. elegans CNS software

rd at sanger.ac.uk rd at sanger.ac.uk
Fri Sep 1 12:50:35 EST 1995

As Leon mentioned, I worked on the wiring diagram data during my PhD.
As part of this I accumulated all the synaptic connectivity
information from the two nerve ring reconstructions done by Eileen
Southgate, John White et al., and also some information on which
neurons are in contact with which, and how much.

All these data are available in two ways.  The connectivity data have
been in acedb for over a year now.  The neurons are objects in the
Cell class, and have a Neurodata entry listing for each other neuron
they connect to the type of connection. and the number of contacts
seen in each reconstruction (N2U and JSH).  Second, the data are also
available in text file version by anonymous ftp from ftp.sanger.ac.uk,
in directory pub/rd, file neurodata.tar.Z.  When you uncompress and
untar this file you get a directory neurodata/ with a README and NOTES
file describing the contents of all the other files.  The data in
these files were compiled from the same source as the figures in
John's paper, and were cross-checked extensively; I would guess that
there are minor differences in detail from Achacoso and Yamamoto.

Unfortunately, I came to the conclusion early on that, although the
worm wiring data are very valuable for those working on the worm
nervous system, they are inadequate as a basis for simulating neural
function.  The programs I wrote as a student looked for motifs in the
circuit diagram and did statistical analysis on the reproducibility of
the connections, rather than simulating them.  Apart from the
possibility of missing connections mentioned by Leon, we are missing a
vast amount of essential physiological data.  In only a few cases do
we know the transmitters involved, and in almost no cases the cell
electrical properties, ion channels etc.  People working on much
simpler circuits, such as the lobster stomach ganglion, where much
more is known experimentally, have trouble enough simulating it.  Of
course, you don't need to simulate the nervous system to get a good
idea of how part of it might work, just as you don't need to simulate
gene regulation pathways to understand them.

Richard Durbin

More information about the Celegans mailing list