# for emacs: -*- mode: sh; -*- # Drosophila virilis -- # # Agencourt's 12 July 2004 assembly # # # DOWNLOAD SEQUENCE (DONE 11/5/04 angie) ssh kksilo mkdir /cluster/store8/droVir1 cd /cluster/data ln -s /cluster/store8/droVir1 droVir1 cd /cluster/data/droVir1 mkdir jkStuff bed mkdir downloads cd downloads wget http://rana.lbl.gov/drosophila/assemblies/dvir_agencourt_arachne_12july04.tar.gz tar xvzf dvir_agencourt_arachne_12july04.tar.gz faSize contigs.fa #189176135 bases (0 N's 189176135 real) in 34715 sequences in 1 files #Total size: mean 5449.4 sd 16095.1 min 90 (contig_34714) max 324616 (contig_3061) median 1204 #N count: mean 0.0 sd 0.0 faSize scaffolds.fa #196642585 bases (7466450 N's 189176135 real) in 27948 sequences in 1 files #Total size: mean 7036.0 sd 145953.7 min 90 (scaffold_27947) max 8062012 (scaffold_0) median 1089 #N count: mean 267.2 sd 6676.2 rm contigs.fa # PARTITION SCAFFOLDS FOR REPEATMASKER RUN (DONE 11/5/04 angie) # Split large scaffolds into 500000-base pieces. # Agglomerate the small scaffolds up into ~500k collections. ssh kksilo cd /cluster/data/droVir1 faSplit size downloads/scaffolds.fa 500000 -oneFile \ scaffoldsSplit -lift=jkStuff/scaffoldsSplit.lft mkdir chunks faSplit about scaffoldsSplit.fa 400000 chunks/chunk_ # RUN REPEAT MASKER (DONE 11/7/04 angie) # January ("March") '04 version of RepeatMasker and libs. # make the run directory, output directory, and job list ssh kksilo cd /cluster/data/droVir1 cat << '_EOF_' > jkStuff/RMDrosophila #!/bin/csh -fe cd $1 /bin/mkdir -p /tmp/droVir1/$2 /bin/cp ../chunks/$2 /tmp/droVir1/$2/ pushd /tmp/droVir1/$2 /cluster/bluearc/RepeatMasker/RepeatMasker -s -spec drosophila $2 popd /bin/cp /tmp/droVir1/$2/$2.out ./ /bin/rm -fr /tmp/droVir1/$2/* /bin/rmdir --ignore-fail-on-non-empty /tmp/droVir1/$2 /bin/rmdir --ignore-fail-on-non-empty /tmp/droVir1 '_EOF_' # << this line makes emacs coloring happy chmod +x jkStuff/RMDrosophila mkdir RMRun RMOut cp /dev/null RMRun/RMJobs foreach f ( chunks/*.fa ) set chunk = $f:t echo ../jkStuff/RMDrosophila \ /cluster/data/droVir1/RMOut $chunk \ '{'check in line+ /cluster/data/droVir1/$f'}' \ '{'check out line+ /cluster/data/droVir1/RMOut/$chunk.out'}' \ >> RMRun/RMJobs end # do the run ssh kk9 cd /cluster/data/droVir1/RMRun para create RMJobs para try, check, push, check,... #Completed: 400 of 400 jobs #Average job time: 2895s 48.25m 0.80h 0.03d #Longest job: 6550s 109.17m 1.82h 0.08d #Submission to last job: 17280s 288.00m 4.80h 0.20d # Lift up the split-scaffold .out's to scaffold .out's ssh kksilo cd /cluster/data/droVir1 foreach f (RMOut/*.fa.out) liftUp $f:r:r.scaf.out jkStuff/scaffoldsSplit.lft warn $f > /dev/null end # Make a consolidated scaffold .out file too: head -3 RMOut/chunk_00.fa.out > RMOut/scaffolds.fa.out foreach f (RMOut/*.scaf.out) tail +4 $f >> RMOut/scaffolds.fa.out end # Load the .out files into the database with: ssh hgwdev hgLoadOut droVir1 /cluster/data/droVir1/RMOut/scaffolds.fa.out # hgLoadOut made a "scaffolds_rmsk" table even with -table=rmsk, # but we want a non-split with no prefix table: hgsql droVir1 -e 'rename table scaffolds_rmsk to rmsk' # Fix up the indices too: hgsql droVir1 -e 'drop index bin on rmsk; \ drop index genoStart on rmsk; \ drop index genoEnd on rmsk; \ create index bin on rmsk (genoName(14), bin); \ create index genoStart on rmsk (genoName(14), genoStart); \ create index genoEnd on rmsk (genoName(14), genoEnd);' # CREATING DATABASE (DONE 11/5/04 angie) # Create the database. ssh hgwdev # Make sure there is at least 5 gig free for the database df -h /var/lib/mysql #/dev/sdc1 1.8T 640G 1021G 39% /var/lib/mysql hgsql '' -e 'create database droVir1' # Copy all the data from the table "grp" # in an existing database to the new database ssh hgwdev hgsql droVir1 -e 'create table grp (PRIMARY KEY(NAME)) select * from hg17.grp' # EXTRACTING GAP INFO FROM BLOCKS OF NS (DONE 11/5/04 angie) ssh kksilo mkdir /cluster/data/droVir1/bed/fakeAgp cd /cluster/data/droVir1/bed/fakeAgp faGapSizes ../../downloads/scaffolds.fa \ -niceSizes=5,10,20,25,30,40,50,100,250,500,1000,10000,100000 # Wow, quite a few blocks of N's are exactly 25bp long. That is # probably the min gap size for the assembly, and it also happens to # be the default value of -minContigGap for hgFakeAgp, so run with it: hgFakeAgp ../../downloads/scaffolds.fa fake.agp ssh hgwdev hgLoadGap -unsplit droVir1 /cluster/data/droVir1/bed/fakeAgp/fake.agp # SIMPLE REPEATS (TRF) (DONE 11/5/04 angie) ssh kksilo mkdir /cluster/data/droVir1/bed/simpleRepeat cd /cluster/data/droVir1/bed/simpleRepeat nice trfBig -trf=/cluster/bin/i386/trf \ ../../downloads/scaffolds.fa \ /dev/null -bedAt=simpleRepeat.bed -tempDir=/tmp \ |& egrep -v '^(Removed|Tandem|Copyright|Loading|Allocating|Initializing|Computing|Scanning|Freeing)' \ > trf.log & # check on this with tail -f trf.log # Load this into the database as so ssh hgwdev hgLoadBed droVir1 simpleRepeat \ /cluster/data/droVir1/bed/simpleRepeat/simpleRepeat.bed \ -sqlTable=$HOME/kent/src/hg/lib/simpleRepeat.sql # FILTER SIMPLE REPEATS (TRF) INTO MASK (DONE 11/5/04 angie) # make a filtered version of the trf output: # keep trf's with period <= 12: ssh kksilo cd /cluster/data/droVir1/bed/simpleRepeat awk '{if ($5 <= 12) print;}' simpleRepeat.bed > trfMask.bed # MASK FA USING REPEATMASKER AND FILTERED TRF FILES (DONE 11/7/04 angie) ssh kksilo cd /cluster/data/droVir1 maskOutFa -soft downloads/scaffolds.fa \ bed/simpleRepeat/trfMask.bed scaffolds.fa maskOutFa -softAdd scaffolds.fa RMOut/scaffolds.fa.out scaffolds.fa # Now clean up the unmasked chunks to avoid confusion later. rm -r chunks scaffoldsSplit.fa jkStuff/scaffoldsSplit.lft # STORE SEQUENCE AND ASSEMBLY INFORMATION (DONE 11/7/04 angie) # Translate to 2bit ssh kksilo cd /cluster/data/droVir1 faToTwoBit scaffolds.fa droVir1.2bit # Make chromInfo.tab. mkdir bed/chromInfo twoBitInfo droVir1.2bit stdout \ | awk '{printf "%s\t%s\t/gbdb/droVir1/droVir1.2bit\n", $1, $2;}' \ > bed/chromInfo/chromInfo.tab # Make symbolic a link from /gbdb/droVir1 to the 2bit. ssh hgwdev mkdir -p /gbdb/droVir1 ln -s /cluster/data/droVir1/droVir1.2bit /gbdb/droVir1/ # Load chromInfo table. hgsql droVir1 < $HOME/kent/src/hg/lib/chromInfo.sql hgsql droVir1 -e 'load data local infile \ "/cluster/data/droVir1/bed/chromInfo/chromInfo.tab" into table chromInfo' # Make chrom.sizes from chromInfo contents and check scaffold count. hgsql droVir1 -N -e 'select chrom,size from chromInfo' \ > /cluster/data/droVir1/chrom.sizes wc -l /cluster/data/droVir1/chrom.sizes # 27948 /cluster/data/droVir1/chrom.sizes # MAKE HGCENTRALTEST ENTRY AND TRACKDB TABLE (DONE 11/5/04 angie) # Warning: genome and organism fields must correspond # with defaultDb values echo 'INSERT INTO dbDb \ (name, description, nibPath, organism, \ defaultPos, active, orderKey, genome, scientificName, \ htmlPath, hgNearOk, hgPbOk, sourceName) values \ ("droVir1", "Jul. 2004", "/gbdb/droVir1", "D. virilis", \ "scaffold_6:5128001-5158000", 1, 57, \ "D. virilis", \ "Drosophila virilis", "/gbdb/droVir1/html/description.html", \ 0, 0, "Agencourt 12 July 2004");' \ | hgsql -h genome-testdb hgcentraltest echo 'INSERT INTO defaultDb (genome, name) values ("D. virilis", "droVir1");' \ | hgsql -h genome-testdb hgcentraltest # Make trackDb table so browser knows what tracks to expect: ssh hgwdev cd ~/kent/src/hg/makeDb/trackDb cvsup . # Edit trackDb/makefile to add droVir1 to the DBS variable. mkdir drosophila/droVir1 # Create a simple drosophila/droVir1/description.html file. cvs add drosophila/droVir1 cvs add drosophila/droVir1/description.html make update DBS=droVir1 ZOO_DBS= # go public on genome-test cvs ci makefile cvs ci drosophila/droVir1 mkdir /gbdb/droVir1/html # in a clean, updated tree's kent/src/hg/makeDb/trackDb: make alpha # PUT SEQUENCE ON /ISCRATCH FOR BLASTZ (DONE 11/5/04 angie) # First, agglomerate small scaffolds into chunks of ~1M median # so we don't have too many files for one dir, but keep a reasonably # low job run time: ssh kksilo cd /cluster/data/droVir1 mkdir chunks faSplit about scaffolds.fa 1000000 chunks/chunk_ ssh kkr1u00 mkdir /iscratch/i/droVir1 cp -pR /cluster/data/droVir1/chunks /iscratch/i/droVir1/ cp -p /cluster/data/droVir1/droVir1.2bit /iscratch/i/droVir1/ iSync # PRODUCING GENSCAN PREDICTIONS (DONE 11/7/04 angie) ssh kksilo # Make hard-masked scaffolds and split up for processing: cd /cluster/data/droVir1 maskOutFa scaffolds.fa hard scaffolds.fa.masked mkdir chunksHardMasked faSplit about scaffolds.fa.masked 1000000 chunksHardMasked/chunk_ mkdir /cluster/data/droVir1/bed/genscan cd /cluster/data/droVir1/bed/genscan # Check out hg3rdParty/genscanlinux to get latest genscan: cvs co hg3rdParty/genscanlinux # Make 3 subdirectories for genscan to put their output files in mkdir gtf pep subopt ls -1S ../../chunksHardMasked/chunk*.fa > chunks.list cat << '_EOF_' > gsub #LOOP gsBig {check in line+ $(path1)} {check out line gtf/$(root1).gtf} -trans={check out line pep/$(root1).pep} -subopt={check out line subopt/$(root1).bed} -exe=hg3rdParty/genscanlinux/genscan -par=hg3rdParty/genscanlinux/HumanIso.smat -tmp=/tmp -window=2400000 #ENDLOOP '_EOF_' # << this line keeps emacs coloring happy gensub2 chunks.list single gsub jobList ssh kki cd /cluster/data/droVir1/bed/genscan para create jobList para try, check, push, check, ... #Completed: 101 of 101 jobs #Average job time: 63s 1.06m 0.02h 0.00d #Longest job: 261s 4.35m 0.07h 0.00d #Submission to last job: 515s 8.58m 0.14h 0.01d # If there are crashes, diagnose with "para problems". # If a job crashes due to genscan running out of memory, re-run it # manually with "-window=1200000" instead of "-window=2400000". # Concatenate scaffold-level results: ssh kksilo cd /cluster/data/droVir1/bed/genscan cat gtf/*.gtf > genscan.gtf cat subopt/*.bed > genscanSubopt.bed cat pep/*.pep > genscan.pep # Clean up: rm -r /cluster/data/droVir1/chunksHardMasked # Load into the database as so: ssh hgwdev cd /cluster/data/droVir1/bed/genscan # Reloaded without -genePredExt 1/6/05: ldHgGene -gtf droVir1 genscan genscan.gtf hgPepPred droVir1 generic genscanPep genscan.pep hgLoadBed droVir1 genscanSubopt genscanSubopt.bed # MYTOUCH FIX - jen - 2006-01-24 sudo mytouch droVir1 genscanPep 0501071300.00 # MAKE DOWNLOADABLE FILES (DONE 11/7/04 angie) # FASTA FIXED 3/21/05 angie, see below ssh kksilo mkdir /cluster/data/droVir1/zips cd /cluster/data/droVir1 zip -j zips/scaffoldOut.zip RMOut/scaffolds.fa.out zip -j zips/scaffoldFa.zip scaffolds.fa zip -j zips/scaffoldFaMasked.zip scaffolds.fa.masked zip -j zips/scaffoldTrf.zip bed/simpleRepeat/trfMask.bed foreach f (zips/*.zip) echo $f unzip -t $f | tail -1 end ssh hgwdev mkdir /usr/local/apache/htdocs/goldenPath/droVir1 cd /usr/local/apache/htdocs/goldenPath/droVir1 mkdir bigZips database # Create README.txt files in bigZips/ and database/ to explain the files. cd bigZips cp -p /cluster/data/droVir1/zips/*.zip . md5sum *.zip > md5sum.txt # SWAP DM1-DROVIR1 BLASTZ (DONE 11/7/04 angie) ssh kksilo mkdir -p /cluster/data/droVir1/bed/blastz.dm1.swap.2004-11-07/axtScaffold ln -s blastz.dm1.swap.2004-11-07 /cluster/data/droVir1/bed/blastz.dm1 cd /cluster/data/droVir1/bed/blastz.dm1 set aliDir = /cluster/data/dm1/bed/blastz.droVir1 cp $aliDir/S1.len S2.len cp $aliDir/S2.len S1.len # With 10k scaffolds, we don't want a directory with one file per # scaffold. So just make one .axt with everything -- not too huge # anyway, given these little insect genomes. zcat $aliDir/axtChrom/chr*.axt.gz \ | axtSwap stdin $aliDir/S1.len $aliDir/S2.len stdout \ | axtSort stdin dm1.axt # CHAIN MELANOGASTER BLASTZ (DONE 11/8/04 angie) # Run axtChain on kolossus (one big dm1.axt input) ssh kolossus mkdir /cluster/data/droVir1/bed/blastz.dm1/axtChain cd /cluster/data/droVir1/bed/blastz.dm1/axtChain axtChain -verbose=0 ../dm1.axt /cluster/data/droVir1/droVir1.2bit \ /cluster/data/dm1/nib stdout \ | chainAntiRepeat /cluster/data/droVir1/droVir1.2bit \ /cluster/data/dm1/nib stdin stdout \ | chainMergeSort stdin > all.chain # Load chains into database ssh hgwdev cd /cluster/data/droVir1/bed/blastz.dm1/axtChain hgLoadChain -tIndex droVir1 chainDm1 all.chain # NET MELANOGASTER BLASTZ (DONE 11/8/04 angie) ssh kksilo cd /cluster/data/droVir1/bed/blastz.dm1/axtChain chainPreNet all.chain ../S1.len ../S2.len stdout \ | chainNet stdin -minSpace=1 ../S1.len ../S2.len stdout /dev/null \ | netSyntenic stdin noClass.net # Add classification info using db tables: ssh hgwdev cd /cluster/data/droVir1/bed/blastz.dm1/axtChain netClass -noAr noClass.net droVir1 dm1 melanogaster.net # Make a 'syntenic' subset: ssh kksilo cd /cluster/data/droVir1/bed/blastz.dm1/axtChain rm noClass.net netFilter -syn melanogaster.net > melanogasterSyn.net # Load the nets into database ssh hgwdev cd /cluster/data/droVir1/bed/blastz.dm1/axtChain netFilter -minGap=10 melanogaster.net | hgLoadNet droVir1 netDm1 stdin netFilter -minGap=10 melanogasterSyn.net \ | hgLoadNet droVir1 netSyntenyDm1 stdin # MAKE AXTNET (DONE 11/8/04 angie) ssh kksilo cd /cluster/data/droVir1/bed/blastz.dm1/axtChain netToAxt melanogaster.net all.chain /cluster/data/droVir1/droVir1.2bit \ /cluster/data/dm1/nib stdout \ | axtSort stdin melanogasterNet.axt # MAKE VSDM1 DOWNLOADABLES (DONE 11/8/04 angie) ssh kksilo cd /cluster/data/droVir1/bed/blastz.dm1/axtChain nice gzip *.{chain,net,axt} ssh hgwdev mkdir /usr/local/apache/htdocs/goldenPath/droVir1/vsDm1 cd /usr/local/apache/htdocs/goldenPath/droVir1/vsDm1 cp -p /cluster/data/droVir1/bed/blastz.dm1/axtChain/all.chain.gz \ melanogaster.chain.gz cp -p /cluster/data/droVir1/bed/blastz.dm1/axtChain/melanogaster.net.gz . cp -p /cluster/data/droVir1/bed/blastz.dm1/axtChain/melanogasterNet.axt.gz . # Make a README.txt which explains the files & formats. md5sum *.gz */*.gz > md5sum.txt # MAKE 11.OOC FILE FOR BLAT (DONE 11/8/04 angie) # Use -repMatch=100 (based on size -- for human we use 1024, and # fly size is ~4.4% of human judging by gapless dm1 genome size from # featureBits -- we would use 45, but bump that up a bit to be more # conservative). ssh kkr1u00 mkdir /cluster/bluearc/droVir1 blat /cluster/data/droVir1/droVir1.2bit /dev/null /dev/null -tileSize=11 \ -makeOoc=/cluster/bluearc/droVir1/11.ooc -repMatch=100 #Wrote 14387 overused 11-mers to /cluster/bluearc/droVir1/11.ooc cp -p /cluster/bluearc/droVir1/*.ooc /iscratch/i/droVir1/ iSync # AUTO UPDATE GENBANK MRNA RUN (DONE 11/22/04 angie) ssh hgwdev # Update genbank config and source in CVS: cd ~/kent/src/hg/makeDb/genbank cvsup . # Edit etc/genbank.conf and add these lines (note scaffold-browser settings): # droVir1 (D. virilis) droVir1.genome = /iscratch/i/droVir1/droVir1.2bit droVir1.mondoTwoBitParts = 1000 droVir1.lift = no droVir1.refseq.mrna.native.load = no droVir1.refseq.mrna.xeno.load = yes droVir1.refseq.mrna.xeno.pslReps = -minCover=0.15 -minAli=0.60 -nearTop=0.005 droVir1.genbank.mrna.xeno.load = yes # GenBank has no D. virilis ESTs at this point... that may change. droVir1.genbank.est.native.load = no droVir1.genbank.est.xeno.load = no droVir1.downloadDir = droVir1 droVir1.perChromTables = no cvs ci etc/genbank.conf # Since D. virilis is a new species for us, edit src/lib/gbGenome.c. # Pick some other browser species, & monkey-see monkey-do. cvs diff src/lib/gbGenome.c make cvs ci src/lib/gbGenome.c # Edit src/align/gbBlat to add /iscratch/i/droVir1/11.ooc cvs diff src/align/gbBlat make cvs ci src/align/gbBlat # Install to /cluster/data/genbank: make install-server ssh eieio cd /cluster/data/genbank # This is an -initial run, (xeno) RefSeq only: nice bin/gbAlignStep -srcDb=refseq -type=mrna -initial droVir1 & tail -f [its logfile] # Load results: ssh hgwdev cd /cluster/data/genbank nice bin/gbDbLoadStep -verbose=1 -drop -initialLoad droVir1 featureBits droVir1 xenoRefGene #15497250 bases of 189178031 (8.192%) in intersection # Clean up: rm -rf work/initial.droVir1 # This is an -initial run, mRNA only: nice bin/gbAlignStep -srcDb=genbank -type=mrna -initial droVir1 & tail -f [its logfile] # Load results: ssh hgwdev cd /cluster/data/genbank nice bin/gbDbLoadStep -verbose=1 -drop -initialLoad droVir1 featureBits droVir1 all_mrna #51271 bases of 189178031 (0.027%) in intersection featureBits droVir1 xenoMrna #15562690 bases of 189178031 (8.226%) in intersection # Clean up: rm -rf work/initial.droVir1 # MAKE GCPERCENT (DONE 11/5/04 angie) ssh hgwdev mkdir /cluster/data/droVir1/bed/gcPercent cd /cluster/data/droVir1/bed/gcPercent # create and load gcPercent table hgGcPercent droVir1 /cluster/data/droVir1 # MAKE HGCENTRALTEST BLATSERVERS ENTRY (DONE 12/?/04 b0b/jen) ssh hgwdev echo 'insert into blatServers values("droVir1", "blat14", "17784", 1, 0); \ insert into blatServers values("droVir1", "blat14", "17785", 0, 1);' \ | hgsql -h genome-testdb hgcentraltest # MAKE Drosophila Proteins track (DONE braney 11/17/04) ssh kksilo mkdir -p /cluster/data/droVir1/blastDb cd /cluster/data/droVir1/blastDb faSplit sequence ../scaffolds.fa 400 x for i in *.fa; do formatdb -i $i -p F 2> /dev/null; done rm *.fa *.log ssh kkr1u00 mkdir -p /iscratch/i/droVir1/blastDb cp /cluster/data/droVir1/blastDb/* /iscratch/i/droVir1/blastDb (iSync) 2>&1 > sync.out mkdir -p /cluster/data/droVir1/bed/tblastn.dm1FB cd /cluster/data/droVir1/bed/tblastn.dm1FB ls -1S /iscratch/i/droVir1/blastDb/*.nsq | sed "s/\.nsq//" > bug.lst exit # back to kksilo cd /cluster/data/droVir1/bed/tblastn.dm1FB mkdir fbfa # calculate a reasonable number of jobs calc `wc /cluster/data/dm1/bed/blat.dm1FB/dm1FB.psl | awk "{print \\\$1}"`/\(150000/`wc bug.lst | awk "{print \\\$1}"`\) # 18735/(150000/396) = 49.460400 split -l 49 /cluster/data/dm1/bed/blat.dm1FB/dm1FB.psl fbfa/fb cd fbfa for i in *; do pslxToFa $i $i.fa; rm $i; done cd .. ls -1S fbfa/*.fa > fb.lst mkdir blastOut for i in `cat fb.lst`; do mkdir blastOut/`basename $i .fa`; done cat << '_EOF_' > blastGsub #LOOP blastSome $(path1) {check in line $(path2)} {check out exists blastOut/$(root2)/q.$(root1).psl } #ENDLOOP '_EOF_' cat << '_EOF_' > blastSome #!/bin/sh BLASTMAT=/iscratch/i/blast/data export BLASTMAT g=`basename $2` f=/tmp/`basename $3`.$g for eVal in 0.01 0.001 0.0001 0.00001 0.000001 1E-09 1E-11 do if /scratch/blast/blastall -M BLOSUM80 -m 0 -F no -e $eVal -p tblastn -d $1 -i $2 -o $f.8 then mv $f.8 $f.1 break; fi done if test -f $f.1 then if /cluster/bin/i386/blastToPsl $f.1 $f.2 then liftUp -nosort -type=".psl" -pslQ -nohead $3.tmp /iscratch/i/dm1/protein.lft warn $f.2 mv $3.tmp $3 rm -f $f.1 $f.2 exit 0 fi fi rm -f $f.1 $f.2 $3.tmp exit 1 '_EOF_' chmod +x blastSome gensub2 bug.lst fb.lst blastGsub blastSpec ssh kk cd /cluster/data/droVir1/bed/tblastn.dm1FB para create blastSpec para try, push # Completed: 151668 of 151668 jobs # CPU time in finished jobs: 4155662s 69261.03m 1154.35h 48.10d 0.132 y # IO & Wait Time: 696054s 11600.90m 193.35h 8.06d 0.022 y # Average job time: 32s 0.53m 0.01h 0.00d # Longest job: 10728s 178.80m 2.98h 0.12d # Submission to last job: 76564s 1276.07m 21.27h 0.89d cat << '_EOF_' > chainGsub #LOOP chainSome $(path1) #ENDLOOP '_EOF_' cat << '_EOF_' > chainSome (cd $1; cat q.*.psl | simpleChain -prot -outPsl -maxGap=25000 stdin ../c.`basename $1`.psl) '_EOF_' chmod +x chainSome ls -1dS `pwd`/blastOut/fb?? > chain.lst gensub2 chain.lst single chainGsub chainSpec para create chainSpec # should run this on the mini-cluster or with my shove script # so you can limit the number of jobs starting to 3 or 4 para try, push... # Completed: 383 of 383 jobs # CPU time in finished jobs: 1220s 20.34m 0.34h 0.01d 0.000 y # IO & Wait Time: 10767s 179.44m 2.99h 0.12d 0.000 y # Average job time: 31s 0.52m 0.01h 0.00d # Longest job: 109s 1.82m 0.03h 0.00d # Submission to last job: 942s 15.70m 0.26h 0.01d exit # back to kksilo cd /cluster/data/droVir1/bed/tblastn.dm1FB/blastOut for i in fb?? do awk "(\$13 - \$12)/\$11 > 0.6 {print}" c.$i.psl > c60.$i.psl sort -rn c60.$i.psl | pslUniq stdin u.$i.psl awk "((\$1 / \$11) ) > 0.60 { print }" c60.$i.psl > m60.$i.psl echo $i done sort -T /tmp -k 14,14 -k 16,16n -k 17,17n u.*.psl m60* | uniq > /cluster/data/droVir1/bed/tblastn.dm1FB/blastDm1FB.psl ssh hgwdev cd /cluster/data/droVir1/bed/tblastn.dm1FB hgLoadPsl droVir1 blastDm1FB.psl # End tblastn # SWAP CHAINS FROM DM2, BUILD NETS ETC. (REDONE 5/23/05 angie) # Originally done 3/11/05 -- redone (better params) 5/23/05. mkdir /cluster/data/droVir1/bed/blastz.dm2.swap cd /cluster/data/droVir1/bed/blastz.dm2.swap doBlastzChainNet.pl -swap /cluster/data/dm2/bed/blastz.droVir1/DEF \ >& do.log & tail -f do.log # Add {chain,net}Dm2 to trackDb.ra if necessary. # FIX X BASES -> N IN DOWNLOADABLE FASTA FILES (DONE 3/21/05 angie) ssh kksilo cd /cluster/data/droVir1 mv scaffolds.fa scaffolds.fa.bak tr 'Xx' 'Nn' < scaffolds.fa.bak > scaffolds.fa diff scaffolds.fa{.bak,} | less # N/n sequence changes> grep '^>' scaffolds.fa.bak > /tmp/1 grep '^>' scaffolds.fa > /tmp/2 diff /tmp/[12] # mv scaffolds.fa.masked scaffolds.fa.masked.bak tr 'Xx' 'Nn' < scaffolds.fa.masked.bak > scaffolds.fa.masked diff scaffolds.fa.masked{.bak,} | less grep '^>' scaffolds.fa.masked.bak > /tmp/1 grep '^>' scaffolds.fa.masked > /tmp/2 diff /tmp/[12] rm zips/scaffoldFa*.zip zip -j zips/scaffoldFa.zip scaffolds.fa zip -j zips/scaffoldFaMasked.zip scaffolds.fa.masked ssh hgwdev cd /usr/local/apache/htdocs/goldenPath/droVir1/bigZips rm scaffoldFa*.zip md5sum.txt ln -s /cluster/data/droVir1/zips/scaffoldFa*.zip . nice md5sum *.zip > md5sum.txt # MAKE Drosophila Proteins track (DONE braney 06-30-05) ssh kk mkdir -p /cluster/data/droVir1/bed/tblastn.dm2FB cd /cluster/data/droVir1/bed/tblastn.dm2FB ls -1S /iscratch/i/droVir1/blastDb/*.nsq | sed "s/\.nsq//" > target.lst mkdir fbfa # calculate a reasonable number of jobs calc `wc /cluster/data/dm2/bed/blat.dm2FB/dm2FB.psl|awk "{print \\\$1}"`/\(80000/`wc target.lst | awk "{print \\\$1}"`\) # 18929/(80000/396) = 93.698550 split -l 94 /cluster/data/dm2/bed/blat.dm2FB/dm2FB.psl fbfa/fb cd fbfa for i in *; do pslxToFa $i $i.fa; rm $i; done cd .. ls -1S fbfa/*.fa > fb.lst mkdir -p /cluster/bluearc/droVir1/bed/tblastn.dm2FB/blastOut ln -s /cluster/bluearc/droVir1/bed/tblastn.dm2FB/blastOut for i in `cat fb.lst`; do mkdir blastOut/`basename $i .fa`; done tcsh cat << '_EOF_' > blastGsub #LOOP blastSome $(path1) {check in line $(path2)} {check out exists blastOut/$(root2)/q.$(root1).psl } #ENDLOOP '_EOF_' cat << '_EOF_' > blastSome #!/bin/sh BLASTMAT=/iscratch/i/blast/data export BLASTMAT g=`basename $2` f=/tmp/`basename $3`.$g for eVal in 0.01 0.001 0.0001 0.00001 0.000001 1E-09 1E-11 do if /scratch/blast/blastall -M BLOSUM80 -m 0 -F no -e $eVal -p tblastn -d $1 -i $2 -o $f.8 then mv $f.8 $f.1 break; fi done if test -f $f.1 then if /cluster/bin/i386/blastToPsl $f.1 $f.2 then liftUp -nosort -type=".psl" -pslQ -nohead $3.tmp /cluster/data/dm2/bed/blat.dm2FB/protein.lft warn $f.2 mv $3.tmp $3 rm -f $f.1 $f.2 $f.3 $f.4 exit 0 fi fi rm -f $f.1 $f.2 $3.tmp $f.3 $f.8 $f.4 exit 1 '_EOF_' chmod +x blastSome gensub2 target.lst fb.lst blastGsub blastSpec para create blastSpec para push # Completed: 79992 of 79992 jobs # CPU time in finished jobs: 1933706s 32228.44m 537.14h 22.38d 0.061 y # IO & Wait Time: 403514s 6725.23m 112.09h 4.67d 0.013 y # Average job time: 29s 0.49m 0.01h 0.00d # Longest finished job: 17598s 293.30m 4.89h 0.20d # Submission to last job: 42181s 703.02m 11.72h 0.49d ssh kki cd /cluster/data/droVir1/bed/tblastn.dm2FB tcsh cat << '_EOF_' > chainGsub #LOOP chainSome $(path1) #ENDLOOP '_EOF_' cat << '_EOF_' > chainSome (cd $1; cat q.*.psl | simpleChain -prot -outPsl -maxGap=25000 stdin ../c.`basename $1`.psl) '_EOF_' chmod +x chainSome ls -1dS `pwd`/blastOut/fb?? > chain.lst gensub2 chain.lst single chainGsub chainSpec para create chainSpec para push # Completed: 202 of 202 jobs # CPU time in finished jobs: 1254s 20.90m 0.35h 0.01d 0.000 y # IO & Wait Time: 6632s 110.53m 1.84h 0.08d 0.000 y # Average job time: 39s 0.65m 0.01h 0.00d # Longest finished job: 218s 3.63m 0.06h 0.00d # Submission to last job: 580s 9.67m 0.16h 0.01d cd /cluster/data/droVir1/bed/tblastn.dm2FB/blastOut for i in fb?? do awk "(\$13 - \$12)/\$11 > 0.6 {print}" c.$i.psl > c60.$i.psl sort -rn c60.$i.psl | pslUniq stdin u.$i.psl awk "((\$1 / \$11) ) > 0.60 { print }" c60.$i.psl > m60.$i.psl echo $i done sort -u -T /tmp -k 14,14 -k 16,16n -k 17,17n u.*.psl m60* > /cluster/data/droVir1/bed/tblastn.dm2FB/blastDm2FB.psl cd .. ssh hgwdev cd /cluster/data/droVir1/bed/tblastn.dm2FB hgLoadPsl droVir1 blastDm2FB.psl exit # back to kksilo rm -rf blastOut # End tblastn # GENEID PREDICTIONS FROM IMIM (DONE 7/26/05 angie) ssh hgwdev mkdir /cluster/data/droVir1/bed/geneid cd /cluster/data/droVir1/bed/geneid wget http://genome.imim.es/genepredictions/D.virilis/golden_path_200407/geneidv1.2/scaffolds.gtf ldHgGene -gtf -genePredExt droVir1 geneid *.gtf