Working Group 7 - Network Basins

Results from the group discussion

Draft version 1, please edit!!!

How to define meaningful edges from CAGE data

• Collaborators are encouraged to provide regulatory interactions in their own way (including clear description of method used)
  • RIKEN will perform MARA analysis similar to FANTOM4
• define a format for the interactions, might include evidence codes, scores, etc.

References

• ARACNe algorithm for TF -> target gene interaction inference (PubMed IDs): 16723010, 17406294, 15778709
• MINDy algorithm for identification of post-translational modulators of TF activity: 15778709, 19741643

How to compare networks

• Network comparison is important for FANTOM5 and will have to be developed
• For a review on network comparison:[ http://mpba.fbk.eu/sites/mpba.fbk.eu/files/visintainer10introduction.pdf Visintainer et al, NIPS2010]

Integration of other datasets

PPI data

Distal regulation

• ChIPing of a few factors will elucidate important enhancers
• factors currently under discussion: P300, H3K4me1, CTCF, H3K27Ac, Pol2
• possible to perform on 10 - 30 samples
• further discussions will clarify in which cells/ time points to perform ChIPing

Gene chains

• collaborators are encouraged to consider gene chains and their relevance to the regulation of the contained genes.

( Specifically, we have a genomewide dataset of human gene chains per Engstrom et al 2006 FANTOM3. Test the hypothesis that chains encode cis-regulatory networks. neighboring genes = nodes. antisense overlaps or bidirectional promoters = edges. do any of these nodes match nodes of conventional trans-regulatory networks that we discussed on 2/22 AM? - LL )

TF and lincRNA in network

• collaborators are encouraged to provide regulatory interactions for lincRNAs

(Note: we already have a genomewide set of lncRNA-mRNA pairs -- antisense pairs Media:F5_human_sense-antisense_pairs_hg19.zip and bidirectional promoters. Also we and others have smaller datasets of well-validated lncRNA-TF regulatory interactions. Let's test the hypothesis that some of these interactions are essential to steady state networks? LL )

Initial focus: lncRNAs that may regulate TFs. We will find " lncRNA gene - TF gene " sense-antisense pairs Media:F5_human_sense-antisense_pairs_hg19.zip in our human sense-antisense lncRNAome annotation; focus on lncRNAs that we have validated by our lncRNAome microarray, e g in human brain; and work specifically on integrating lncRNA-TF CIS-regulation with TF-to-targetGene TRANS-regulation! - LL

How to (and do we) integrate miRNAs in the network

• expression of mature miRNAs together with public miRNA-_target_ predictions can be employed to predict regulatory interactions (miRNA and mRNA = nodes, negative regulatory interaction = edge).

MiRNAs sometimes act outside of 3'UTRs. - LL