Choice of gene sets analysed for athropod gene structure are based on 
expert decision on best available gene sets at the time of analysis.
Rather provide analyses of all species in a related group, a best
representative is chosen.  For Drosoophila, the model D. melanogaster
has far more EST and experimental evidence than other species. For
mosquitoes, the choice is less obvious.

Genome annotations are more like computers than they are like
vintage wines, i.e. the newer models are generally better, they
do not improve as much with age.

Use of Culex for a representative mosquito is based on its newer,
more extensive gene annotation pipeline, and large EST data set
coupled with a robust EST assembly/integration into gene
set.  This choice is also supported by the higher average identity
of Culex genes to other arthropod gene sets.

Newer gene models are using improved software, and more software
programs, which has been shown to improve results over fewer
(gene finding errors average out).  There is also a hard to
quantify effect of conservation for older genome annotations.
Older genome annotations are updated from time to time, but
experience says there is a degree of resistance to large changes
suggested by new annotation piplelines.  Rather new automatic
annotations are merged with older ones in a conservative way.

- Don Gilbert, 2009 Sept., 

Mosquito gene sets used by euGenes/Arthropods (2008 August)

Culex quinquefasciatus   ESTs 205275
  Genome sequencing status: 04/19/2007, draft assembly 
  available gene set:    Dec  2008, cpipiens.Cpip1.2
  Gene finding quality report
  http://www.broadinstitute.org/annotation/genome/culex_pipiens.4/GeneFinding.html
  
Anopheles gambiae        ESTs 153269
  Genome sequencing status: 03/22/2002, draft assembly 
  available gene set:    Dec  2008, agambiae.AgamP3.4
  
Aedes aegypti            ESTs  301596
  Genome sequencing status: 02/11/2005, draft assembly 
  available gene set:    Jan  2008, aaegypti.AaegL1.1.gff3

Culex gene finding has these improvements to other species:
  1. newer and more gene predictor methods (Augustus, Snap, PHAT, GeneID, FGeneSH)
  2. newer EST assembly methods (PASA,nap)
  3. Guidance from prior species gene models
  

Gene set quality can also be measured from average protein
identity to 11 other arthropod gene sets. 
            
            ngroup  ave-PI se-PI
Aedes	    	10664	  42.76  0.21
Anopheles	   9106	  35.41  0.23
Culex	    	11038	  44.27  0.20

Data source: http://insects.eugenes.org/arthropods/orthologs/
Total gene groups: 12582

The highest identity is found for Culex, with Aedes a close
second.  This can reflect taxonomy, but also the computational
gene prediction quality when measured on full gene sets of these
related species.


Culex annot notes:

http://cpipiens.vectorbase.org/Genome/GeneView/?gene=CPIJ012099;db=core

Genes were annotated by merging VectorBase, JCVI and BROAD
predictions sets. VectorBase automatic analysis pipeline using
either a GeneWise/Exonerate model from a database protein or a
set of aligned cDNAs/ESTs followed by an ORF prediction.
GeneWise/Exonerate models are further combined with available
aligned cDNAs/ESTs to annotate UTRs (For more information see
V.Curwen et al., Genome Res. 2004 14:942-50). JCVI annotations
were modeled based on a combination of Augustus (M.Stanke),
GlimmerHMM (M.Pertea), Twinscan (M.Brent), Snap (I.Korf), and
PHAT (S.Cawley) coupled with protein and EST genome alignments
(PASA (M.Campbell), Genewise (E.Birney) and nap (X.Huang)).BROAD
annotations were generated based on a combination of Augustus,
GeneID (R.Guigo), FGeneSH (A.Salamov), GeneWise and SNAP and
completed by species specific protein alignments (BLAST).

Anogam annot notes:

http://agambiae.vectorbase.org/Genome/GeneView/?gene=AGAP009325;db=core
Genes were annotated by merging Ensembl and TIGR predictions
sets. Ensembl automatic analysis pipeline using either a
GeneWise/Exonerate model from a database protein or a set of
aligned cDNAs/ESTs followed by an ORF prediction.
GeneWise/Exonerate models are further combined with available
aligned cDNAs/ESTs to annotate UTRs (For more information see
V.Curwen et al., Genome Res. 2004 14:942-50). TIGR annotations
were modeled based on a combination of Augustus (M. Stanke),
GlimmerHMM (M. Pertea), Genie (D. Kulp), and Twinscan (M. Brent)
gene predictions, coupled with protein and EST genome alignments.

Aedes annot notes:

http://aaegypti.vectorbase.org/Genome/GeneView/?gene=AAEL000078
Genes were annotated by merging Ensembl and TIGR predictions
sets. Ensembl automatic analysis pipeline using either a
GeneWise/Exonerate model from a database protein or a set of
aligned cDNAs/ESTs followed by an ORF prediction.
GeneWise/Exonerate models are further combined with available
aligned cDNAs/ESTs to annotate UTRs (For more information see
V.Curwen et al., Genome Res. 2004 14:942-50). TIGR annotations
were modeled based on a combination of Augustus (M. Stanke),
GlimmerHMM (M. Pertea), Genie (D. Kulp), and Twinscan (M. Brent)
gene predictions, coupled with protein and EST genome alignments.