Dear maize biologists,
I've reconstructed an accurate maize gene set with EvidentialGene,
using 3 recent Illumina RNA-seq sets, as version evg5corn here:
This set is more accurate by homology to other plant species, than three
other recent maize gene sets. Don't take my word; the measurements are
objective, so have a look at it, and let me know what you think.
Genes were reconstructed using de-novo RNA assembly, without
using chromosomes. I did however need to use chromosome mapping to
help separate the many high identity loci (ohnologs/homeologs) of maize.
Gene reconstruction with EvidentialGene methods are reliably producing
more accurate and complete gene sets for many plants and animals that I
and others have tried it with. Should you be interested in such, please
look into this.
My knowledge of maize genes is limited. I picked on your favorite for
other reasons, the hope that Gramene's PacBio RNA gene reconstruction
would be made public and allow an objective comparison of that method
with Illumina RNA reconstruction methods. For some reason they left
that out of their recent paper (doi:10.1038/ncomms11708).
I substituted Gramene's MAKER built gene models of Zea_mays.AGPv4 from
this August, which appear to have used those PacBio genes, in this
comparison. By objective measures with homology to other plants, it is
not as complete as what I reconstructed with Illumina-RNAseq and
EvidentialGene. Also compared are maize gene sets B73_RefGen_v3 by NCBI,
and genes assembled with Rnnotator by JGI (using same RNA).
-- Don Gilbert