Ques t ions to be answered

1. Questions to be answered Group defined

2. Elements Group • Comparison between target lists for peak‐callers • Simple statistics of binding sites • No. of sites associated with promoters, genes … • Motif analysis for sequence specific factors • Saturation Analysis for Binding Sites • No. of sites identified as a function of number of cell lines (for Pol II & CTCF) • No. of sites identified as a function of the number of TFs (GM12878 or K562)

3. Large Scale Behaviour • Segmentations • How do we convince ourselves that a given segmentation is good? • How do we select an appropriate number of labels and a good model topology? • How do we assign semantics to labels? • How do we evaluate whether a hypothesized label corresponds to a true biological phenomenon? • How can we use the output from other analysis groups to validate and interpret our segmentations?

4. Large Scale Behaviour • Segmentations • What can we say about the similarities and differences of segmentations across cell lines? • Are there specific genomic territories (CNSs, exons, promoters, enhancers, repeat classes, etc.) that we should be focusing on? • How can we make our existing segmentations and analysis tools more useful to members of the ENCODE consortium?

5. Small Scale Integration • Are the different dimensionality reductions saying the same thing? How much of that is technical vs biological? • To what extent is this a QC on segmentation vs other facets to be explored? • What annotation is best to bring along side this dimensionality reduction • Is RNA best used as annotation (“output”) or integrated into the system?

6. Variation Group • Overlap AS Calls for different experiments. • Run data over common pipeline to call AS events. • Use Elements group as way to call peaks in unified way.

7. Questions to be answered Hypothesis defined

8. Integrative analysis of transcriptional units • Canonical transcribed regions. • Comprehensively delineate transcribed regions by combining GENCODE annotations, RNA output, and chromatin modifications. James, Felix, Julian, Nathan, Steve, Bronwen RNA group • Promoters • Delineate active and poised promoters by combining GENCODE annotations, experimental RNA data, polymerase binding, and chromatin accessibility/modification. Piero, Ali, Anschul, Orion Elements • Ditto site of transcript termination

9. Integrative analysis of transcriptional units • Novel/alternative gene units. • Delineate comprehensively gene-like transcribed regions that fall outside of GENCODE annotations by integrating RNA output and chromatin modifications. Mike M., Julian, Ajish, RNA

10. Integrative analysis of transcriptional units • Transcriptional domains. • What proportion of the transcriptional output of the genome is organized into well-defined, multi-gene/transcript domains that can be systematically connected with chromatin modifications, accessibility, and interactions (5C)? Michael H., Ross, Georgi, Large Scale Behaviour. • Are there large-scale chromatin phenomena that are not connected with the transcriptional output of the underlying genomic sequence (e.g., expression status of genes within some defined domain)? Jason, Tony, James, Large Scale Behaviour

11. Transcription vs TFs/Chromatin • Sites of transcript origination. • How much of RNA-based TSS annotation is associated with chromatin/TF features? Promoter question – see earlier. • Whether chromatin/regulatory features can be used to predict these regions accurately in the absence of RNA data? Promoter question – see earlier. • What is the genomic distribution of non-chromatin-associated transcriptional activity? RNA + Elements + Long Range Behaviour?

12. Chromatin • Can we determine the overlap of the DNase1 methods, and DNAse 1 with FAIRE. Terry, Bob, Zhuzhu Elements • What DNAse 1 elements and signal should be used inthe analyses e.g. the element matrix?

13. Transcription vs TFs/Chromatin • Transcription associated with likely cis-regulatory regions. • Many well-annotated cis-regulatory regions (e.g., the beta-globin locus control region) are occupied by polymerase and appear to harbor transcriptional activity. How common is this finding, and does it have predictive potential for a sub-class of RNA transcripts and their compartmentalization? Laura, Brandon, Barbara Elements

14. TFs and Chromatin • Can we identify: • TFs that do not require or correlate with open chromatin. • TFs that correlate with open chromatin BECAUSE they trigger nucleosome remodeling. • TFs that correlate with open chromatin BECAUSE they can only bind accessible DNA. • DNase/FAIRE sites that don't match up to any histone modification or TF binding site. • Punctate histone marks (e.g., H3K4me1) that don't match up with DNase/FAIRE sites. • Alan, Noam, Raymond, Zhiping, Mike Elements

15. CTCF • Can we identify: • CTCF sites at boundaries of domains i.e. Insulators • CTCF sites that lie between regulatory elements and promoters i.e. enhancer blocker sites • CTCF sites at promoters • Jason, James, Tony, Large Scale Behaviour

16. Variation • Allelic variation in transcription. • Determine what proportion of allelic transcription of both canonical and alternative transcribed regions is accompanied by synchronous allelic changes in chromatin/regulatory factor state. • Determine the potential of allelic changes in chromatin/regulatory factor state for predicting allelic transcription of overlapping or nearby genes. • Ewan, Tim, Jay, Preti Variation

17. Comparative • Define the patterns of evolutionary constraint and human genetic variation associated with: • Small Elements Joel, Javier Elements • Co-annotated (RNA+chromatin) features e.g. TSSs, transcript termination sites, others. RNA+Elements+ Comparative • What is the conservation pattern of chromatin-defined domains that are not connected with transcription? Large Scale + Comparative • For DNA segments containing each feature, what fraction of the nucleotides in each functional class show evidence of constraint? What (if any) fraction show evidence of accelerated evolution? Joel, Javier Elements