MAExplorer - Microarray Exploratory Data Analysis

Glossary of terms used in MAExplorer

Abbreviations:
- E.C.L. is Edited Gene list
- G.C. is Gene Class
- H.P. is hybridized sample
- HP-E is list of expression profile hybridized samples
- HP-X is list of X-axis hybridized samples
- HP-Y is list of Y-axis hybridized samples
Applet - a java program that runs in a Web browser and is downloaded from the Web server each time it is run. Applets do not require installing any software on your computer. However, they may not read or write any files on your computer. MAExplorer was initially used used as an applet when accessing microarray data on the original MGAP http://www.lecb.ncifcrf.gov/maemicroarray Web server.
Background - the background intensity measured near a spot in a microarray image. Spot intensity measurements from images with high background or where background varies over the image because of non-specific hybridization in parts of the array need to be corrected. Some data sets have this background available as a single average value for the entire array and others may have it on a per-spot bases. MAExplorer can optionally subtract this background if it is available.
cDNA - complementary DNA immobilized as spots on the target microarray are hybridized with mRNA of the hybridized sample. Alternative methods of constructing arrays attach synthetic oligonucleotides to the arrays rather than cDNA from clones.
Clone - consisting of many copies of a particular cDNA. Each spot in the microarray target represents cDNA from a different clone. In MAExplorer we use the term gene to refer to both clone spotted arrays and oligonucliotide arrays.
Clone ID - The unique identifier assigned to the clone by the I.M.A.G.E. consortium. This may be used to lookup the clone and related data in the NCBI dbEST, GenBank, UniGene, NCI/CIT mAdb clone databases. The Clone ID is reported in all results from MAExplorer searches.
Cluster - a set of genes with similar hybridized samples expression profiles. It may also be a set of HPs with similar expression profiles for a specific set of genes. Different clustering methods define "similarity" in different ways. See Cluster Plots in the Reference Manual.
Condition - is a named set of samples defined or edited using one of the condition choosers. One is the (Samples menu | Choose HP-X, HP-Y and HP-E samples) to define the current working sets. The other defines a named condition set using the (Samples menu | Choose named condition lists of samples). Condition sets are used in various operations including clustering and statistical tests. For example the t-Test compares the HP-X and HP-Y condition sets of replicates.
Current gene - the particular gene being analyzed in MAExplorer. This is set by the user by clicking on a gene in the array, scatter plot or Clone ID instance in a report or typing the name or Clone ID in the popup text window. It is indicated with a green circle in the array and scatter plots, and by a darkened background in the Reports.
Current cluster - the particular subset of genes belonging to a cluster of genes found using the K-means clustering method. These genes belong to the cluster to which the current gene belongs. Therefore, to change the current cluster - change the current gene. If K-means clustering is active and the current cluster is defined, the genes which belong to it will be indicated with a tiny green cluster number in the array image and in the scatter plot (if one is present). When the current cluster is defined, the genes in the current cluster are copied to the 'edited gene list'.
Current condition - the last named condition set edited with the condition chooser using the (Samples menu | Choose named condition lists of samples). It is available for any analysis method that may want to use a particular condition.
dbEST - Expressed Sequence Tag (EST) data from the NCBI dbEST database. The database may be indexed by clone Id and may have entries for the 3' and/or 5' sequenced genes.
DRYROT error - is a fatal error that is detected by MAExplorer, it will popup an error reporting window. We call this a "DRYROT" error (thanks to "S.A.I.L.") because something is wrong in the program or data files and from which it can not recover. This type of error should not have happened.
Duplicate genes - (notation) genes that are spotted multiple times (i.e. F1, F2) for all of the genes on the same array. This is different than replicated genes that may not have copies for all genes.
Edited Gene List - (E.G.L) is a subset of genes in the MAExplorer database. It may be defined and edited manually or may be set from various MAExplorer operations (e.g. set to the current cluster when doing similarity clustering, K-means clustering, etc).
EST - Expressed Sequence Tag of a particular gene fragment which is expressed in a particular tissue.
Expression Profile - (or EP) is the vector of the spot intensity values or expression profile for the same gene for the list of hybridized samples being analyzed. This is used in computing similarity between genes. Alternatively, it may be a vector of gene intensities for a hybridized sample. This is used in computing similarity between samples.
F1,F2 - are the left (F1) and right (F1) replicate fields of the microarray (if the array supports it). Some arrays duplicate the gene in corresponding grids of spots in the microarray (Research Genetics, NIA neuroarrays, etc.). This lets us compute a coefficient of variation (CV) for the gene. If the CV is low, then the measurement is more reliable.
Filter - (data filter) is a sequence of restrictions on the set of all genes used to determine a subset list of genes. These restrictions include gene membership in a "Gene Class" (eg. oncogenes); quantitative tests including: gene class membership, spot-CV-value, ratio or intensity range or membership in histogram bins, statistical and clustering tests, etc. These are then used to pre-filter the set of all genes to a working subset which may then be used for viewing, reporting, or saving in a gene subset. The subset may then be used in subsequent MAExplorer operations or saved for later use.
Gene - in MAExplorer we use the generic term gene to refer to data for both gene spotted arrays and oligonucliotide arrays.
Gene Class - (G.C.) the sets of particular genes with similar properties, e.g. oncogenes, heat shock, milk proteins, etc. These can be used to help partition the set of all genes as part of the data filtering process. The MAExplorer Gene Name Guesser (Section 2.3.1)
Gene expression profile - (or EP) the vector of normalized quantitative values for the corresponding gene in the set of hybridized sample (HP-E). This vector represents the expression of the gene across a set of experimental conditions such as developmental stages, dose-response, time-series, etc.
Gene subset - a set of genes defined by the current filter may be saved and used at a later time as part of the Filter. In addition, there are a number of set operations (union, intersection and difference) which may be used to derive new sets of genes.
Grid - a rectangular matrix of spots on the microarray. Generally the spots are laid out in a series of grids (also called "blocks" or a "patches"). They are indexed within the grid as a "grid row" and a "grid column" If the grids are replicated, they are indicated by multiple fields of duplicated spots.
Grid coordinates - The genes are layed down in a microarray in rows and columns of a grid. There are multiple grids which may have a space between them. For example, in the MGAP microarray database there are 8 grids (named A through H) to a field with a space between grids. Finally, there may be two fields (left and right named 1 and 2 or F1 and F2) which are duplicates of the grids so as to allow us to get an estimate of the hybridized sample variance by spotting the same gene multiple times in the array.
Hierarchical clustering is a clustering method that constructs a binary tree where the leaf objects are the set of genes (or conditions) being clustered. The algorithm is described in Section 3.2.2.3 describing the K-means clustering. It is invoked by the Hierarchical clustering menu entry in Section 2.4.5.4.
Hybridized Sample - (or HP) a particular hybridized sample's quantified data. It may be a hybridized microarray of cDNA clones (or alternatively oligonucleotides) are hybridized with ³³P radio-labled or Cy3 and Cy5 flourescent-labeled mRNA total sample probe. We abbreviate this hybrized sample in MAExplorer as H.P. or HP.
HP-X - the individual H.P. sample assigned to the Y-axis when doing an analysis. The H.P-Y 'set' is a set of samples used to compute mean intensity for each gene.
HP-Y - the individual H.P. sample assigned to the Y-axis when doing an analysis. The H.P-Y 'set' is a set of samples used to compute mean intensity for each gene.
HP-E - an ordered list of hybridized samples that is used in computing the gene expression profile.
Image - the microarray image associated with a hybridized sample with a microarray obtained by scanning the radioactive or flourescent labeled array on a phosphoimager or densitometer (if ³³P labeling was used to expose film) at about 50 or 100 microns/pixel. [see particular brands and models for details.] The image is then quantified to obtain individual spot intensity values using an image analysis program (the image analysis step is not part of MAExplorer).
I.M.A.G.E. - I.M.A.G.E. clone consortium is a public collection of sequenced clones which offers the ability to locate and acquire specific cDNA clones by assigned ID numbers. MAExplorer uses the Clone ID as the primary gene identifier.
Intensity - refers to the integrated density over a spot and corresponds to the level of that hybridized sample's labeled mRNA to the genes cDNA immobilized on the array. Also see background density.
K-means is a clustering method that assigns data Filtered genes to a fixed number of orthogonal clusters. The algorithm is described in Section 3.2.2.2 describing the K-means clustering. It is invoked by the K-means clustering menu entry in Section 2.4.5.3.
Microarray - a nylon membrane, glass slide, or other subtrate with thousands of cDNA clones or oligonucleotides "spotted" in unique positions on the array. The cDNA are typically PCR products of approximately 0.6-2.4kb representing specific genes and typically 100-500 microgram/ml (Duggan, Nature Genetics Supplement, 21, Jan, 1999, pg 11). For example, the initial MGAP database array created by Research Genetics is a nylon membrane (3x7 cm) with 3500 spots (1500 duplicated clones in cDNA and EST inserts). The array may be "hybridized" with labeled samples to quantify how much of each cDNA is present in the sample. See (Schulze, 2001) for a nice summary of the current technology.
mAdb - the NCI/CIT MicroArray Database Program [http://nciarray.nci.nih.gov/] is a repository of microarray data from the NCI ATC (Advanced Technology Center) microarray spotting facility.
mAdb Clone DB - a clone database integrating hypertext links to various genomic databases for a clone. This is part of the mAdb.
MGAP - Mammary Genome Anatomy Program. MAExplorer was initially developed for this Program. This is part of the Biology of the Mammary Gland [http://mammary.nih.gov/] Web site.
Normalization - the method for scaling all spot quantitation measurements on a set of hybridized samples to the same relative range so as that data from the differerent samples may be compared.
Ordered Condition List (OCL) of condition lists - is a named list of conditions defined or edited using the (Samples menu | Choose ordered lists of conditions).This is an list of multiple conditions that you have previously defined. The OCL may be sorted if you want and the data lends itself to sorting. E.g., a time series of conditions lends itself to sorting - different types of diagnoses may not. Ordered Condition Lists are used in various operations such as statistical tests that require conditions sets of replicate samples.
Probe - (MAExplorer convention) total sample labled mRNAs used to test hybridization with target cDNA immobilized on the array. Synonymous with hybridized sample in MAExplorer.
Pseudoarray image - a MAExplorer image representing the quantified spot data in the database. It may or may not correspond to the actual layout of grids, grid-rows and grid-columns in the original array. It is useful for getting the gestalt of the overall changes in the data. There are many representations of the data (see Section 2.4.4.1 Show Microarray for discussion and examples of these representations.
Quantification - the analysis of a scanned microarray image for the purpose of quantifying each spot on the array. This process typically consists of first finding the extent of a spot and then integrating the intensity. If multiple dyes are used (e.g. Cy3, Cy5) then the process is performed for each color-filtered image.
Ratio - in the context of microarrays and MAExplorer, consists of computing the ratio of Cy3/Cy5 or HP-X/HP-Y sets of samples. You need to see the particular context of its use to determine which meaning is implied.
Replicate genes - (notation) genes that are spotted multiple times for some of the genes on the array. This is different than duplicated genes that do have copies for all genes.
Reports - are popup window reports of ordered sets of genes meeting various search criteria may be generated. These consist of the quantitative measure, Clone Id, dbEST id, GenBank Id, Unigene link, etc. These are created in popup windows in MAExplorer as either scrollable dynamic spreadsheets or tab-delimited text windows which can be exported by cutting and pasting into Excel. The default dynamic spreadsheet option, contains active hyperlink fields which point to other web databases (such as GeneCard, dbEST, GenBank, Unigene, mAdb clones, etc.).
Sample - same as hybridized sample in the context of MAExplorer.
SaveAs - is the command or button used in many of the plot windows to save the current plot as a full resolution GIF file specified by the user in a popup file browser window. It is available only in stand-alone mode and not in the applet version. These files could then be used for documenting a data-mining experiment or for publication.
Silhouette plot - is way to graphically display a sorted list of similar objects. In the way it is used in MAExplorer, each object in the descending sorted list (e.g. of genes) has a line (or set of "******") proportional to the similarity to the object at the top of the list. It has the largest number of stars since its similarity is 1.0. See (Kaufman and Rousseeuw, 1990) for discussion on silhouette plots. It may be visualized in the following example as:
```
   Object  Graphic  Similarity
   A       *******   1.00
   B       *****     0.83
   C       *****     0.80
   D       **        0.30
   E       *         0.17
   F                 0.06
```
Similarity measure - is a scalar measure computed from two vectors of normalized measurements for two objects indicating how similar these two objects are. In MAExplorer, the vector is the expression profile (see Section 3.2.1) of the same gene for the list of hybridized samples being analyzed. It may alternatively be the expression profile of a list of genes for a given hybridized sample. So similarity measures indicate how similar two different expression profiles are. In MAExplorer, you may use either an Euclidian distance or correlation-coefficient as the scalar metric.
Similarity cluster is a clustering method that finds the subset of objects (eg. genes or conditions) that is most similar to the selected object. The list of genes is reduced to those whose "distance" between the selected object and the candidate genes is less than some threshold. The algorithm is described in Section 3.2.2.1 describing the similarity clustering. It is invoked by the similarity cluster menu entry in Section 2.4.5.1.
Spot - the spot in a microarray image corresponds to a clone or oligonucleotide attached to the array substrate. The intensity of the spot corresponds to the level of hybridized sample for that gene to the DNA in the spot which correlates to mRNA transcription expression.
Startup file - when using MAExplorer as a stand-alone application, this is the file (with the file name ending with a ".mae" extension) that invokes MAExplorer. On systems where you launch applications by double clicking on them, you would double click on this file. On systems such as UNIX where you specify arguments to a command line interface, you would type "MAExplorer startup .mae file name".
Target - (MAExplorer convention) cDNA or oligonucleotide of a gene immobilized as spot on a microarray. The array contains thousands of these hybridized sample spots.
UniGene - the NCBI UniGene database "is an experimental system group for automatically partitioning GenBank sequences into a non-redundant set of gene-oriented clusters. Each UniGene cluster contains sequences that represent a unique gene, as well as related information such as the tissue types in which the gene has been expressed and map location."
Zdiff - statistical measure of difference between two samples (eg. HP-X and HP-Y) when using the Zscore or Zscore of logs of the data normalizations. If samples x,y have Zscore values for gene c in samples HP-X and HP-Y, or HP-X 'sets' and HP-Y 'sets', then the Zdiff is computed as:
```
   Zdiff(x,y,c) = Zscore(x,c) - Zscore(y,c)
```
Zscore - HP data normalization scaling method to set the mean value to 0.0 and the maximum and minimum values of the scaled values to about +3.0 to -3.0 (standard deviations). It only makes sense if the data is roughly a gaussian distribution. If it is not, you could use the Zscore of the logs of the data to make it more gaussian. If c is a gene in hybridized sample p with mean_p and stdDev_p, then it is computed as:
```
    Zscore(p,c) = (c - mean_p)/stdDev_p
```