!blastx \
-query /Volumes/web/scaphapoda/Grace/Transcriptomes/M_donacium_fasta.fa \
-db /Volumes/Data/blast_db/uniprot_sprot \
-max_target_seqs 1 \
-max_hsps 1 \
-outfmt 6 \
-num_threads 16 \
-out /Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird.tab

!blastx

!tail /Volumes/web/scaphapoda/Grace/Transcriptomes/M_donacium_fasta.fa

!fgrep -c ">" /Volumes/web/scaphapoda/Grace/Transcriptomes/M_donacium_fasta.fa

!tail /Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird.tab

time !wc -l /Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird.tab 

#finished!!

!head /Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID.tab

#translate pipes to tabs

!tr '|' "\t" </Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID.tab> /Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID2.tab

!head /Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID2.tab

#translate commas to tabs

!tr ',' "\t" </Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID2.tab> /Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID3.tab
!head /Volumes/web/scaphapoda/Grace/meodesma_blastx_uniprot_hummingbirdSPID3.tab

#run in SQL share against SPID and GO Numbers and GO_to_GOslim

#"where aspect like 'P'" narrows information down to proteins that are involved in biological processes
#"Column1, term, GOSlim_bin, aspect" select specific columns of information 

!head /Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird_GOSlimterms.tab

#to get unique terms 

!awk '{print $4}' /Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird_GOSlimterms.tab | sort | uniq

#grep all terms pertaining to reproduction/sex-determination/etc

!egrep -wi --color 'aldosterone|female|gametogenesis|genitalia|germ|germ-line|germarium-derived|germinal|germination|hormone|juvenile|male|mating-type|oocyte|ovulation|penile|postnatal|prostate|prostatic|reproduction|vulval' /Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird_GOSlimterms.tab

#want to make chart separating aspects of reproduction
#not sure how to do that

from pandas import *

jslim = read_table("/Volumes/web/scaphapoda/Grace/mesodesma_blastx_uniprot_hummingbird_GOSlimterms.tab", # name of the data file
            #sep=",", # what character separates each column?
            na_values=["", " "]) # what values should be considered "blank" values?

jslim.groupby('GOSlim_bin').Column1.count().plot(kind='bar')