Subject:
Summary of pseudogene call on Thurs. 27-Oct at 11 AM EDT + Overall Group Summary [PGENE]
Date:
Mon, 07 Nov 2005 08:47:43 -0500

Hi,

Deyou and I won't be with you in Santa Cruz. I hope the workshop goes
well!  If people are interested, we may be able to call in on Fri.

Just to make sure everyone is on the same page, I'm sending around the
current list of encode pseudogenes again and am attaching below
minutes from the 27-Oct call plus a overall summary of the group
(which I did for Tom and Roderic).

cheers, marK

##
## Summary of 27-Oct-05 Call
##

1. We are moving towards a very well annotated pseudogene dataset for
ENCODE regions.

2.  We are very interested in transcribed pseudogenes and
understanding if there is any activity associated with them.

3. To answer question number 2, feature integration of various sorts
will give a better idea of how strong the evidence for pseudogene
transcription is.

4. Need to think about what is the best way to intersect pseudogene
and tag sequence data, as tags are usually in UTR's and pseudogene
coordinates will not have those. 
5. Roderic, Tom and Mark will decide on how to go about including
pseudogene representation in the workshop next week.

==

Identification of candidate pseudogenes that could be potentially
transcribed

France used a 1 nucleotide overlap between TAR and pseudogene
coordinates as the criteria for identifying potentially transcribed
pseudogenes. For ditags, only the content on the 5' and 3' ends were
considered. Here again, France used the coordinates of 5' and 3' tag
to pull out pseudogenes which had at least 1bp overlap with the ditag
coordinates

Approximately 50% of pseudogenes could be potentially transcribed. But
a similar analysis also shows similar results for gene
transcription.We expect to not see too many overlaps this way as
pseudogene coordinates are entirely based on protein coding homology
and does not include UTRs.  *Thus, 5% overlap with CAGE tags is
intriguing.

To ponder: *What is a good way to intersect pseudogenes with CAGE and
ditag data?  France suggested looking at the distribution of lengths
of the tags away from pseudogene coordinates and use an optimal number
for thresholding.  i.e. include all pseudogenes that are within 'x'
base-pairs away from a tag sequence as a candidate for potentially
transcribed pseudogene.  Singapore person proposed that genomic
features such as EST data, binding site evidence, CAGE, DITAG,
transfrags..all of these can be integrated to see if it makes a
stronger case for a pseudogene in terms of being potentially
transcribed.

It would also be of interest to check if pseudogenes that could be
potentially transcribed have high sequence identity to the parent
gene.  Presumably, this could mean that it is an artefactual result
due to cross hybridization issues.

==

OVERALL SUMMARY OF SUBGROUP (provided to Tom and Roderic)

We have carefully identified and annotated 198 pseudogenes in ENCODE
regions.  Pseudogenes were identified using different methods by the
various groups and categorized as processed or non-processed
pseudogenes.  All the candidate pseudogenes were merged and a
consensus list of pseudogenes has been obtained. Pseudogenes that were
not identified by all the groups have been carefully inspected
manually and merged with the consensus list after thorough evaluation
of various pieces of evidence such as GENCODE gene annotations and EST
data.  Pseudogenes that are potentially transcribed have been
elucidated by intersecting pseudogene locations on the genome with
transcriptional evidence from TARS obtained by expression studies by
Affymetrix and Yale groups, EST evidence, CAGE and DITAGS.  RACE
experiments are currently underway in 12 human tissues and multiple
RACEs will be pooled to be analyzed by hybridization to Affymetrix
ENCODE 20 nucleotide resolution arrays to experimentally verify
potentially transcribed pseudogenes. The current consensus list of
pseudogenes are integrated in the UCSC browser and also available from
http://www.pseudogene.org/ENCODE/.