Ranked List of All Annotation-based Modules with Significant GO terms |
This page describes all the annotation-based modules in the cisRED Mouse 3.1 database associated with at least one GeneOntology (GO) term with a false discovery rate (FDR) below 0.01, and is divided into three sections:
- Modules overview
information about the modules found in this database - Module details
details about each module in this database, including scores, group composition and more. - Definition of terms
the types of objects contained in this database, and how these objects relate to each other.
Overview of modules in Mouse 3.1
| Parameters used for module discovery | |
| Maximum allowable annotation p-value | 0.0004 |
|---|---|
| Maximum allowable discovery p-value | 0.005 |
| Window width | 300 bp |
| Minimum distance between motif starts | 10 bp |
| Module must be found in at least this many search regions | 10 |
| Summary statistics for modules in this database | |
| Total number of annotation-based modules in the database | 1,600 |
|---|---|
| Total number of annotation-based module instances | 35,047 |
| Total number of search regions that contain annotation-based module instances | 1,573 |
| GO term FDR threshold | 0.01 |
| Total number of GO terms with FDR < FDR threshold | 121 |
| Total number of modules with at least one GO term with FDR < FDR threshold | 537 |
Module details and rank
Each row in the table below represents one module in this database, and the smaller its rank, the better the module.
The rank is determined by first sorting the modules based on its score, and since smaller scores are better, the scores are sorted in ascending order.
If two or more modules have the same score, they are further ranked based on the number of module instances, and since more instances are better, this secondary subranking is done in decreasing order.
In the table below, 'Num. inst.' is the number of module instances. The size of a module is reflected by the number of group IDs listed. Some of these IDs may be duplicates, because a single group can participate in a module multiple times.
For each promoter-based search region that contains an instance of a module, we have found all the Gene Ontology (GO) terms associated with that region, and then, for each of these terms, we calculated the false discovery rate (FDR) for that module. GO terms with a FDR below the FDR threshold of 0.01 are listed in the 'GO terms' column.
- Click on a module ID to visit a page with more information about that module, including a list of all atomic motifs that participate in the module.
- Click on a group ID to visit a page with information about that group.
- Click on a GO term to read more about the term using AmiGO.
- Click on one of the two buttons located directly above the search results to download a tab-delimited file containing the entire set of search results, or just the set of results currently shown on this page.
NOTE: The cisRED modules scoring proceedure is currently under review. At present, all annotation-based modules in this database have the same score, 1.00E-03.
Definition of terms
Each cisRED database is based on a target genome, which is the genome of a target species, such as human, mouse, rat, or c.elegans.
The cisRED pipeline begins with the motif discovery stage, which operates on a large number of independent input sequence sets. An input sequence set consists of a sequence on the target genome, called the search region, and corresponding sequences from many other species. Search regions are generally 2-3k in size and are often located near a gene's transcription start site.
An atomic motif (or motif), is a set (or 'stack') of conserved sequences from some or all species of an input sequence set. A motif target sequence is the sequence that the target genome contributes to a motif. A motif can be represented by its sequence set/stack, or can be represented in a simplified form as a consensus sequence, a position frequency matrix, or a sequence logo.
A motif group (or group) is a set of similar motif target sequences. cisRED identifies groups using two different approaches: annotation and 'de novo'. Each motif group is labeled with a unique ID number. Currently, we consider annotation-based groups and modules to be more reliable than 'de novo' groups and modules. Work to improve genome-scale 'de novo' grouping is ongoing.
'De novo' groups of similar motif target sequences were identified by calculating edit distance dissimilarities between pairs of site sequences, for all significant site sequences, then hierarchically clustering the resulting similarity matrix using the OPTICS algorithm. An atomic motif is assigned to zero or one 'de novo' group.
Annotation-based groups of similar motif target sequences were identified by annotating against known site sequences (rather than against alignment matrices) in TRANSFAC, JASPAR, and ORegAnno. An atomic motif can belong to any number of annotation-based groups.
A module or pattern is a set of co-occuring motif group labels. A module is called either 'annotation-based' or 'de novo' depending on the type of groups it contains. (Currently, all groups in a module must be of the same type.) The size of a module is the number of co-occurring groups in that module. More precisely, a module may contain one or more references to a group label, and each group label reference counts towards the size of the group. A module of size N is called an N-way module.
A module instance (or pattern instance) is a list of motifs, all found in the same search region and in close proximity to each other, in which the motif target sequence of each motif in this list has been assigned to be in at least one of the module's group labels.
The module discovery process is based on a number of parameters, including the size of the 'window' (measured in bp), within which all the motifs in a module instance must be found, and the minimum number of search regions that must contain instances of the module.