HCO-db database
We stored our
simulation files into an archive
and the characteristics of the simulated instances into a database
(HCO-db). This page describes the contents of the HCO-db database of 10.5
million HCO model instances. The construction and analysis of the database
are described in detail in
Doloc-Mihu A, Calabrese RL (2011). A database of
computational models of a half-center oscillator for analyzing how neuronal
parameters influence network activity. J Biol Phys,
Springer, 37(3): 263-283 [PubMed].
·
HCO archive
·
HCO-db database
HCO archive
We stored our
simulation data (files) into an archive. For fast access to these files, we
organized the archive into folders and sub-folders. There are 100 folders
called sim1 to sim100. Each of them (except the last one) contains 100
subfolders named sim_1 to sim_100. Each subfolder contains all the files
corresponding to 1,050 simulated instances.
Each simulated
instance has generated between 4 to 6 files, depending on the activity
type. The names of these files start with the name of the respective
instance, followed by file content information. On average, the files of a
simulated instance occupied about 6MB space on disk.
To save storage space, for each simulated
instance we recorded its voltage traces and the conductances
of ISynS, Ih,
and ICaS of both neurons in a binary
file (a binary files has a smaller size than an ASCII file); further, we
used a disk_out format to be able to compress it.
For compressing the binary (voltage) files, we used Flac
(Free Lossless Audio Codec). The total size of the entire archive of the
files of our 10,485,760 simulated instances was 2.4TB.
The archive is freely available for interested
researchers. Because of its size, it is not practicable to download the archive
over the internet. Instead, we are happy to send a set of DVDs that include
a copy of the archive to anybody who is interested upon e-mail request to
Anca Doloc-Mihu (adolocm at emory.edu).
HCO-db database
We built an efficient relational MySQL database table (HCO-db) with the resulting HCO
instances characteristics. To build the database and to record the
characteristics and the parameters of each simulated model instance into
this database, we used our own Java 1.5 scripts.
For each simulated instance, the script read
the characteristics of the instance from its ASCII file and recorded them
into a row of the database. Each model in the database has its own unique
identifier (unique number which also is a primary key in the database),
which makes it easier to query. In addition for efficient querying, we used
another 13 indexes: for each parameter, for each neuron type, and for the
system type (group label).
We chose the optimal data type for each field
in the database to minimize the storage space required and to speed up the
querying process. For example, the parameter values weren't recorded as
floating point values, but as small integers which gave the percents of the
parameter values with respect to their canonical values. Also, we did not
record the spike files and the voltage files into the database; we just
recorded a pointer to the place in the archive where they reside (the name
of the folder/subfolder for each model). The recorded data occupied approximately
2 GB and the indexes occupied 1 GB; our database has a total size of
approximately 3.06 GB. Click here to see the
structure of HCO-db.
The HCO-db is freely available for interested
researchers. Download the HCO-db files here. (coming soon!)
Last updated July
15, 2012. Please send comments to adolocm@emory.edu.
|