TF network prediction: Difference between revisions
(+overview) |
(general info on the method) |
||
| Line 1: | Line 1: | ||
TF regulatory network prediction |
|||
[[Peter Pemberton-Ross]], Universität Basel, Switzerland |
|||
peter.pemberton-ross@unibas.ch |
|||
=Overview= |
=Overview= |
||
This analysis takes the results from MARA and uses them to construct a |
This analysis takes the results from [[MARA]] and uses them to construct a dynamical model of how TFs/miRNAs affect each others' activities. The output can be visualised as a network. |
||
=Technical description= |
|||
The algorithm steps through a timecourse and looks for constant, reproducible relationships between the predicted activities of TFs at successive timepoints. |
|||
Relationships between the activities which are present throughout the timecourse (non-transient interactions) will be highest-scored, and the confidence of predictions will increase with increasing timecourse length and with more replicates. Relationships which are approximately linear on the sampling timescale are detected best. |
|||
=Application= |
|||
This analysis will be applied to all FANTOM5 timecourses shortly. The resulting network picture will be added to each timecourse's page, and the predicted interactions between pairs of TFs will be made available. |
|||
=Presentation= |
|||
For more detailed information, please see the following teleconference presentation: [[File:Presentation Peter PR.pdf]] |
|||
Revision as of 02:02, 8 November 2012
TF regulatory network prediction
Peter Pemberton-Ross, Universität Basel, Switzerland peter.pemberton-ross@unibas.ch
Overview
This analysis takes the results from MARA and uses them to construct a dynamical model of how TFs/miRNAs affect each others' activities. The output can be visualised as a network.
Technical description
The algorithm steps through a timecourse and looks for constant, reproducible relationships between the predicted activities of TFs at successive timepoints.
Relationships between the activities which are present throughout the timecourse (non-transient interactions) will be highest-scored, and the confidence of predictions will increase with increasing timecourse length and with more replicates. Relationships which are approximately linear on the sampling timescale are detected best.
Application
This analysis will be applied to all FANTOM5 timecourses shortly. The resulting network picture will be added to each timecourse's page, and the predicted interactions between pairs of TFs will be made available.
Presentation
For more detailed information, please see the following teleconference presentation: File:Presentation Peter PR.pdf
