The attributes that are defined for each table depend on the attribute definition for a particular database. A default set of
attributes is supplied which are suitable for capturing all of the
information required for compliance with the MIAME standard. The collection of attributes
can be customised for
The Experiment Section
The Experiment table records information common to a set of Hybridisations such as the goal of
the experiment, the factors examined, a brief description and details
of any publication which describe the experiment.
Each Experiment has a link to a Submitter which identifies the person who
contributed this Experiment to the database.
A Submitter is the person or organisation who supplied one or
Each Measurement represents the collection of results from
analysing the scanned image of the microarray after Hybridisation. Information stored in this
table includes the name of the person who performed the hybridisation,
when it was done and which HybridisationProtocol was used.
A Measurement is linked to the Image
from which the results were obtained. More than one Measurement
can be linked to a particular Image if, for
example, the Image was scanned multiple times
using a different ScanningProtocol.
Measurements can be linked to any number of
Property instances in
which the actual experimental data is stored. Each Property
corresponds to one column in the file that came from the scanner, or
to data derived in subsequent data analysis such as normalisation.
The Image records details of the image file that is produced
when an Array has been analysed by a
scanner. Relevant details include which ScanningProtocol was used, who the scan
was done by, when it was done and the resolution and colour depth of
the resulting image.
The image data itself is not stored in the database. Images should
be archived elsewhere and a link to where the archived image should be
recorded in the Image instance.
Hybridisation is the process in which one or more LabelledExtracts are placed on an
Array which is then treated according to a given HybridisationProtocol. Afterwards,
the Array is scanned, resulting in an Image which is subsequently analysed to acquire the
expression profile data.
An Array repesents a single microarray 'chip' which is used
to perform a Hybridisation. The
Array table records data including the batch number of the
chip and the date on which it was manufactured.
Each Array must be linked to an ArrayType which contains the actual description
of the array design used to manufacture the array.
The Array Design Section
ArrayType records information that is common to a set of
Array objects, i.e. the
description of a particular design of microarray. Each Array that uses this design should be
linked to this ArrayType instance.
Each ArrayType can be linked to by any number of Feature instances, each of which represents
a single well or spot on the microarray described by the
The suggested form for ArrayType names is "Lab Name - Species Name -
Unique Identifier", for example "Bloggs - Homo Sapiens - xyz200/v2".
A Feature represents a unique location on an Array. Each Feature has an ArrayType link recording which ArrayType the Feature belongs to. A
Feature can only belong to one ArrayType.
Each Feature normally has a link to a Reporter, however this link can be omitted if it
is not known what has been spotted or synthesised in this
Features record the location of the spot (both as an absolute
position on the array and as a row/column location) and whether this
Feature is a control element (and if so, what type of
A Reporter (sometimes refered to as a 'probe') is a nucleotide
sequence that has been placed in one or more Features on an Array.
Reporter can be linked to any number of Genes. These links identify what the Reporter
is supposed to be detecting. Multiple Reporters can be linked
to the same Gene in the case where several short sequences
are used to identify the presence of a single larger
sequence. Alternatively, a single Reporter can be linked to
more than one Gene when it is known that several
sequences are likely to bind to the same material.
Reporters can be shared between ArrayTypes. If the same material is used to
create several different ArrayTypes then the same collection of
Reporters can be used when creating the Features.
Each Reporter instance stores information including the
'type' of the Reporter (e.g. cDNA clone or plasmid), the
sequence data, references to external genomic databases and how this
Reporter was manufactured.
The Gene is a general purpose term refering to any
nucleotide sequence who's presence can be detected by one or more
A Gene instance stores information including the
'type' of the Gene (e.g. EST or exon), the
actual sequence data and any references to external genomic databases
providing further information.
Each Gene can be linked to by any number of Reporters.
The Bio-Materials Section
A Source represents the organism, tissue sample or individual
from which all the other bio-materials are derived. It is the starting
point for a series of processes which result in a LabelledExtract that can be used for Hybridisation.
The Source tables can record a great deal of information, not
all of which is relevant to every domain. Possible data that can be
captured includes the organism name, the sex, age and developmental
stage of the individual, details of the cell type or line used, and
information about any known disease carried by the Source.
Where possible, terminology used to describe the Source
should come from well known ontologies (i.e. standard
dictionaries of well-defined terms, for example the NCBI
Taxonomy). More information about providing terms from ontologies can
be found here.
A Sample represents the acquistion of material from a Source. The process by which the Sample is
obtained is described using a SamplingProtocol.
Information that can be stored about a Sample includes the
date on which it was created, the name of the person who did the
sampling and the type of material obtained.
The Sample is one suitable point for use in recording the
underlying experimental design. In a "treated vs. untreated" style
experiment for example, two Samples might be taken from the same
A TreatedSample is a Sample which has
had something done to it. Details of the particular treatment
performed are recorded in the associated TreatmentProtocol.
The TreatedSample table is unique in that it allows a
recursive link, i.e. a TreatedSample can be linked to
another TreatedSample. This enables arbitrarily long chains of
TreatedSamples to be created, with each step in the chain
representing the application of a different treatment. An example of
this might be TreatedSample 'X1' was dosed with some compound,
resulting in TreatedSample 'X2' which was then incubated for 24
hours, resulting in TreatedSample 'X3' which was then dosed
with a different compound resulting in TreatedSample
An Extract is the final stage in bio-material preparation
prior to the creation of a LabelledExtract.
The particular extraction method that was used is recorded in the
ExtractionProtocol link. The
Extract table can also be used to record when the extraction
event occured and the name of person who performed it.
An Extract is linked to one or more TreatedSamples. Linking to multiple TreatedSamples enables a 'pooling' operation
to be represented, in which material from several TreatedSamples is combined to provide
sufficient material for Hybridisation.
A LabelledExtract is an Extract which has
been combined with a compound that is detectable by a scanner,
typically a fluorescent dye or radioactive marker.
Details of how the Extract was labelled are
recorded in the LabellingProtocol
which is linked to each LabelledExtract.
Information about when the labelling was done and who did it can be
stored in this table. The quantity of material that was produced must
also be recorded.
The Protocols Section
The protocol tables record information about a specific task
performed as part of the experiment. Typically a given laboratory will
have a small number of standardised protocols that are reused between
experiments. However, variations from these standard protocols should
always be captured by creating a new protocol instance and linking to
that rather than the standard one from which it is derived.
All of the protocol tables share a common format, storing
a description of the equipment (hardware and/or software) that was
used and a reference to a publication or other resource which
describes in protocol in more detail.
The SamplingProtocol describes the process of preparing a Sample of material from a specific Source.
A TreatmentProtocol captures information about something
that was done to a Sample (or TreatedSample) as part of the
experiment. Examples might include "exposure to radiation" or
"incubation for a time period".
The ExtractionProtocol details how material was extracted
from one or more TreatedSamples in
preparation for labelling.
A LabellingProtocol records the steps taken to attach some
form of detectable marker to an Extract so that
the expression levels of candidate sequences can subsequently be
quantified in the scanning process.
The HybridisationProtocol should provide full details of how
the Hybridisation was performed,
including the blocking agent(s) used, the wash procedure and all
relevant parameters settings of the hybridisation station.
The ScanningProtocol describes the hardware and/or software
used to create an Image by scanning the Array after the Hybridisation was completed.
The ScanningProtocol should record all of the parameters of the
scanning hardware (for example laser power, voltages and gain
An ImageAnalysisProtocol describes the process by which an
Image is analysed to generate a series of data
which can be stored in the corresponding Measurement. Full details of the software
used, and the parameters chosen should be provided.