Geoduck Gonad Gene Annotations

After kicking around how to make a very big table with all of the annotations… I finally made some progress.

tldr; Annotation table with SwissProt, Gigaton, Ruphibase, Dheilly-comparison.

todo: need to add sex specific expression.


I started with a list of all 153982 “contigs”.

01_2_1-Geo-Trans-MasterTable_1D95D66A.png

After getting quite fare with join I eventually leveraged SQLShare. Here are a few bash moments I was fond of…

!join -t $'\t' -a 1 \
analyses/Geoduck-transcriptome-v3-IDonly.tab \
analyses/Geo-v3-join-uniprot-all0916-condensed.sorted \
> analyses/_Master-SP

outputs tab and keeps all of file 1 - important is that -t $'\t' needs to go before arguments.

I cleaned up a blast table

!sort analyses/Geoduck-v3_blastn_GIGAton_e10.out | awk -v OFS='\t' '{ print $1, $11, $2}' \
> analyses/Geoduck-v3_blastn_GIGAton_e10.out.sorted

and with -v OFS='\t' made sure output was tab-delimited.

you can even read in and out on different delimiters…

!awk -F, -v OFS='\t' '{print $2,$11,$13,$14,$20,$21}' analyses/Dheilldy-Gametogenesis-matches.csv

What crushed me was if there was no SP match - the next join went directly adjacent.

01_2_1-Geo-Trans-MasterTable_1D95D986.png


Then I went to SQLShare

The first join was with the SP blast (I could use the bashed up files to upload to SQLShare).

01_2-Geo-Trans-annotation_1D95D620.png

SELECT
*
FROM [roberts.sbr@gmail.com].[Geoduck-transcriptome-v3-IDonly]ID
left join
[roberts.sbr@gmail.com].[Geo-v3-join-uniprot-all0916]sp
on
ID.Column1 = sp.[GeoID]
left join
[roberts.sbr@gmail.com].[Geoduck-v3_blastn_GIGAton_e10]gig
on
ID.Column1 = gig.Column1
left join
[roberts.sbr@gmail.com].[Geoduck-v3_blastn_RuphiBase_e10]ru
on
ID.Column1 = ru.Column1
left join
[roberts.sbr@gmail.com].[Dheilly-comparison-Geo-v3]Dh
on
ID.Column1 = Dh.GEOID

I downloaded this… then !open analyses/_Master-SP-Gig-Ruphi_Dh.csv -a "Microsoft Excel" and succumbed to Excel to clean up a bit and validate.

_Master-SP-Gig-Ruphi_Dh_1D95DB9F.png

Now each row is a sequence (contig) a 158k row table.

I think all we need now is to add expression data!


edit

Written on September 23, 2016