Daphnia New tile array expression predicted genes, 2008 February, D. Gilbert

Data at http://microbe.bio.indiana.edu:7182/prerelease2/gene-predictions/
with methods at http://microbe.bio.indiana.edu:7182/data/dpx-augtrials/

Summary

We have genome tile array expression that has exon/intron level
resolution.  I'ved used those data with Augustus gene prediction
as coding hints,' to call the genes that all the predictors are
missing, but that have good expression signals.  What I get that
way is about 30% additional CDS sequence, although the gene
models produced with tile expression hints are not as good as
other ones (shorter).

Using tile expression as CDS hints recovers most, but not all, of
the expression region hints (depending on weights) but results in
short/partial gene models.  Poorer gene models than no hints, and
increasing weight to force predictions reduces gene model
quality.

# Augustus with TAR, augmap19.gff.gz, all scaffolds
CDSbases aug19   :  ntr=56928,  n=197323,  m=214.95,  cds=42413860,  tb=162548203,   c/t=0.261
   ( CDS overlapping JGI genes: 29511951,  no-overlap JGI genes:  12901909 ; 42% additional)

#  Daphnia v1.1 genome gene set coding sequence bases / total genome bases
CDSbases JGI_V11 :  ntr=30940   n=142754,  m=211.28,  cds=30160786,  tb=174233412,   c/t=0.173
CDSbases Gnomon  :  ntr=37466   n=151668,  m=237.45,  cds=36014074,  tb=200738384,   c/t=0.179

Key: ntr: number transcripts; n:number CDS-exons; m:mean exon size; 
     cds:cds bases, tb:total gene-region genome bases, c/t: cds/total ratio

Using Daphnia tile expression data adds ~ 40% of new CDS,  e.g.  daphnia  c/t=0.26
John Manak's study (Nature genetics, 2007, doi:10.1038/ng1875) w/ Drosmel 
tile array expression suggests 30%  transcription outside predicted genes, 
e.g.  drosmel  c/t=0.24 versus drosmel known genes c/t=0.18

Note however this study found only a small portion in intergenic regions. 
Due to daphnia's short introns and early gene prediction set, most of Daphnia's
unfound genes are between current gene predictions.

I also matched these new TAR predicted genes to known genes, and to
Daphnia ESTs.

Only 10% of Dpx AugTar19 proteins are found in NCBI-nr database (eval<1e-3)
EST support 12% of the new AugTar19 genes.
      
Top homology species

311	 Insects
 175	 Nasonia + Apis
  82	   Tribolium	
  50	   Aedes + Drosophila	 
215	 Aquatic (Zebrafish, Sea urchin, anemone)
238	 Bacteria
132	 Transposon genes
577  Other 

311	 Insects
 108     Nasonia		
  82     Tribolium	
  65     Apis				
  32     Aedes			
  24     Drosophila	 

215	 Aquatic
  93     Zebrafish      (Danio)
  90     Sea urchin    (Strongylocentrotus)
  32     Sea anemonae		(Nematostella   (sea anemonae)

132  Transposon genes
	
238	 Bacteria
  59     Acidovorax    (soil bacteria)
  51     Verminephrobacter (bacteria)
  43     Pseudomonas       (bacteria)
  34     Methylibium    (soil bacteria)
  26     Delftia				(bacteria)
  25     Rhodoferax		(bacteria)

577 Other 

Top Homology Descriptions
Count	Name
262	Hypothetical protein
49	Transposase 
19	transcriptional regulator
11	Peptidase
11	unnamed protein product 
9	Orf2-encoded protein
8	Orf2-encoded protein 
7	major facilitator superfamily mfs_1 
7	orf2-encoded protein 
6	Osjnba0011f23.1
6	ankyrin repeat protein
5	Lreo_3 
5	abc transporter related 
4	Conserved Hypothetical protein
4	Of unknown function duf6
4	abc membrane transporter 
4	heavy metal translocating p-type atpase 
4	inner-membrane translocator 
4	tonb-dependent siderophore receptor 
4	two component transcriptional regulator
  ..

#---------------------