|wFleaBase | BLAST | BioMart | GBrowse Maps | Genomics | Help|
2007: Genes 1.0 (JGI_V11) Genome DE|
Showing differential expression partitioned into Gene, Intron or Unknown regions of genome, as percent of genome.
2010: Genes 2.0 (Best3) Genome DE|
Showing differential expression partitioned into Gene (CDS), UTR, Intron or Unknown regions of genome, as percent of genome.
Expression MA density plots for Genes 2.0
These density plots show the relative frequency of
differential expression (M value, Y axis)
versus absolute expression intensity (A value, X axis). They
help understand that the rather large DE in UTR regions
is consistent across treatment types, and associated with
somewhat higher absolute expression.
|Sex Coding Genic||Sex UTR||Metal Coding Genic||Metal UTR|
|Cadmium Coding Genic||Cadmium UTR||Chaoborus Coding Genic||Chaoborus UTR|
Six tandem Chorion peroxidase genes are better resolved in Genes 2.0.
Two on right (genes 5,6) show NO Male, and weak Chaoborus expression.
One on left (gene 1) shows stronger +Male expression, and strong Juvenile (Cha/Cha-con).
Genes 3,4 in middle show +Male equal or more than Female, and Gene 4
also shows +Chaborus.
This annotated list dpxmix19_tilede_effects.txt
has some of the obvious DE cases, selected as the strongest DE (M value) from
new genes and UTR/ncRNA regions in Cadmium, Chaborus and Mixed Metals.
See also this shorter dpxmix19_tilede_interesting.txt
list of about 40 most interested new genes and ncRNA regions for Cadmium,
Mixed metals and Chaborus. Maps of expression for some of these interesting cases are shown
in folder dpxmix19_tilede_interestpix/.
Genes with any DE on CDS, all treatments DE JGI Gnomon Best3 Change (b3-jgi) cad+ 148 130 146 -2 cad- 84 56 76 -10 cha+ 438 389 562 +130 cha- 567 530 579 +10 met+ 696 574 1376 +650 met- 2223 2064 2363 +100 sexf 4196 4089 4662 +450 sexm 2894 2569 3454 +550 nul 23226 30273 38779
Notable stats for Best3 vs JGI_V11 genes: -- Best3 recovers 82% of protein homology, versus only 42% for JGI -- Best3 recovers 90% of EST, versus 75% for JGI -- Best3 recovers 66% (5Mb) of 7.6 Mb tile expression not in JGI set (tar_genes) Best3/beta3 (dpulex_aug26_mixin19h): 48060 mRNA in mixin19h set, no alt-transcripts yet 42893 mRNA have evidence (homology, est, parology or tile expression) 5167 mRNA have no evidence, but protein >= 40aa minimum 5661 mRNA are non-coding by weak cds criteria (protein <40aa or <30% coding) 22717 have protein homology (e-value <= 1e-5) 18681 have EST evidence 12409 have differential expression 6079 have tar-gene expression (outside JGI set) 10693 have protein paralogy only (e-value <= 1e-5)Map views are at http://server7.wfleabase.org/cgi-bin/gbrowse/daphnia_pulex8/
At this directory,
this folder matab19_1004/
has tables of expression A and M (DE) values, and DE significance (p-value) for 4 treatments (cad, cha, met, sex) and five feature types:
1. gene : coding sequence of Genes 2.0, 2. exonutr : untranslated regions of Genes 2.0, including UTR of coding genes and ncRNA 3. intron : of genes 2.0 4. taru : TAR regions not covered by Genes 2.0 5. unann : regions with expression (A) not in an of the above features. Tables are in this format ID part ntiles A M tmean ndf pmean hxAUG25p1s1g100t1 utr 54 1.94029 0.12018 0.46454 95 0.64332 hxAUG25p1s1g101t1 utr 39 1.60615 0.06057 0.27037 65 0.78773 hxAUG25p1s1g103t1 utr 105 0.49133 -0.04401 -0.15340 167 0.87826For Cadmium, see these tables (I think you can ignore the intron, unann ones):
See also the MA plots showing density of these effects, e.g. introns have little A or M, while gene and exonutr have A up to 6 or so (log scale) and M in -2 to 2 range.
I sorted these tables by highest M value and pulled out the top 20 or so to annotate. E.g. for +Cad in new coding genes not in JGI gene set
gzcat gene.txmastat104.cad.tab.gz | perl -ne\ 's/(\d\.\d+e\-\d+)/0/g; ($g,$tp,$nt,$a,$m,$tm,$nd,$p)=split; \ $am=sprintf "%.3f", $a - abs($m)/2; s/\-?[\d\.]+\t(\d+)\t([\d\.]+)$/$am\t$2/; print if ($nt>8);'\ | sort -k5,5nr | ggrep -v -F -f $dpxe/evda/mixin19.jgifull.ids - | head -20For +Cad in UTR regions
gzcat exonutr.txmastat104.cad.tab.gz | perl -ne\ 's/(\d\.\d+e\-\d+)/0/g; ($g,$tp,$nt,$a,$m,$tm,$nd,$p)=split; \ $am=sprintf "%.3f", $a - abs($m)/2; s/\-?[\d\.]+\t(\d+)\t([\d\.]+)$/$am\t$2/; print if ($nt>4);'\ | sort -k5,5nr | head -20MS Excel should be able to sort results like this also. You will find analogous tile DE results for the JGI gene set (1.1) here:
Analysis methods used for tile expression are outlined here http://server7.wfleabase.org/prerelease2/dpxtiles0907/dpxtilegene09-recipe.txt