Daphnia magna gene set 201304 status
2014 Jan, from Don Gilbert, gilbertd at indiana.edu

NOTE: Dmagna5x_evigene_2014jan is a pre-release gene set that will be
replaced in summer 2014.  
 
0. Prior D. magna gene sets of 2011, genome-modelled, and 2012, RNA-seq
assembled are incomplete (missing ortholog families, short/fragmented
genes).  DMagna project folks can find these at
http://server7.wfleabase.org/genome/Daphnia_magna/prerelease/gene-predictions/
  
1. 2013 gene set is constructed from StressFlea + NotreDame mRNA-seq
with EvidentialGene methods.  It follows current RNA-seq de-novo
assembly methods, adding 20% genome-modelled genes for low expression,
orthology models. This gene set has 38,800 loci with 144,000 alternate 
transcripts. There is 35% orthology improvement over the 2011 gene set,
and ~90% are fully supported by expression.

2. Orthology and expression validations are essentially done.  This D. magna gene
set has more complete genes (longer) than D. pulex, and about same number of
ortholog gene families.  [ picture here.. ]
This D. magna gene set has a bit more Human-orthology than D. pulex,
and more than any other tested arthropod among 9 insects, 4 arachnids, 2 shrimp.
    
3. The major remaining problem is accurately classifying among
paralogs, alternate transcripts and artifactual assemblies.  It is both
a hard problem, and an important one to solve.  E.g., I have not yet been
able to accurately count and model all the Hemoglobin genes in Daphnia
magna, primarily because they are high identity recent paralogs. These
are also differentially expressed genes, responding to several
environmental stressors.  This problem holds for many gene families.

4.  I'm for looking collaborators to publish this gene set, sooner rather
than later this year.  It appears better quality than current D. pulex
genes in useful respects, and adds a needed good quality crustacean gene
set.  A limitation is my pressing need to find funding to support this
work.

Strong human-gene orthology, coupled with existing methods of measuring
environmentally responsive genes, reinforce the role for Daphnia as an
important model organism for enviro-genomics.  Having a very accurate
Daphnia gene set for this is an important step that I ask your help
with.


*  "The quality of a gene set is dependent on the quality of the genome assembly" 

This dogma is now wrong, a quote from one and implicit in many recent
genome projects. With today's accurate and inexpensive RNA-seq data, the
quality of eukaryote gene sets is more related to quality of mRNA-gene assembly
(e.g. http://arthropods.eugenes.org/EvidentialGene/about/EvidentialGene_quality.html)

Complete and biologically accurate expressed gene catalogs can now be determined
better from mRNA seq assembly than from gene prediction on genome assemblies.
Using both approaches improves genes, but also turns up conflicting gene evidence
that makes reaching a higher level of accuracy a difficult problem.

See http://arthropods.eugenes.org/EvidentialGene/about/alntopbars/index.html
for comparison of gene set quality for insects, crustacea, ticks, fish and
plants. In all comparable cases, mRNA-assemblies have more complete orthology
genes than genome-modelled genes.


* Eco-Environ relevant genes remain a hard problem.

- Recently evolved genes are the hardest to accurately assembly or
model, with duplication repeat problems, weak or no orthology models,
variable expression (i.e. lower than orthology genes for standard environs,
but high for special environs). Such are subject to mis-assembly,
gene-joins, poor genome mapping, gapping, and other gene construction
problems.

- Environmentally responsive genes are more often recently evolved,
species-specific or new paralogs, with lesser known functions inferred
from orthology. This finding turned up for Daphnia pulex, and appears
with Daphnia magna and Killifish, as well as other gene x environ
studies.

Daphnia magna Gene Orthology x Expression effects
  Express Ortholog Inparalog Unique
    HS7     9%-      12%     78%+  Cyanobacteria,Carbaryl,Crowding,Fish,Parasite,Triops/HS
    ND4    22%-      20%     58%+  Cadmium,Salt,pH5,UVlight/ND
    NDxPB  50%+      10%-    40%   Lead/ND
  No diff  31%       23%     46%   no differential expression
------------------------------------------------------
  Express = High expression group gene proportions for
    HS=StressFlea/Helsinki, ND=NotreDame/Pfrender
  Ortholog  = has ortholog gene outside Daphnia (insect, arthro, vert)
  Unique    = species-specific genes
  Inparalog = Inparalogs (new genes), including Daphnia magna+pulex
------------------------------------------------------

Killifish Gene Orthology x Expression effects
   Express    Inparalog  Unique  
    Adult-Env   31%+     35%+  
    Embryo       7%-     11%-  
    No diff     12%      31%   
------------------------------------------------------
  Express = High expression group gene proportions for
    Adult-Env = from adult tissues, with environ stressors
    Embryo = from embryonic tissues, no stressors
  Unique  = species-specific genes versus ortholog genes
  Inparalog = Inparalogs (new genes) versus orthologs
------------------------------------------------------

My opinion on accurate reconstruction of recent-genes is that available
gene data is often sufficient, but improved methods are needed for
construction and validation.

- Developing new methods for inferring or measuring gene functions of these
environmentally responsive, recently evolved genes is an important
problem for researchers in this area now (Daphnia-philes and others).