src/hg/makeDb/doc/dasNov1.txt 1.7
1.7 2009/07/21 21:01:41 markd
added transmap update blurb
Index: src/hg/makeDb/doc/dasNov1.txt
===================================================================
RCS file: /projects/compbio/cvsroot/kent/src/hg/makeDb/doc/dasNov1.txt,v
retrieving revision 1.6
retrieving revision 1.7
diff -b -B -U 1000000 -r1.6 -r1.7
--- src/hg/makeDb/doc/dasNov1.txt 4 Dec 2008 00:59:07 -0000 1.6
+++ src/hg/makeDb/doc/dasNov1.txt 21 Jul 2009 21:01:41 -0000 1.7
@@ -1,846 +1,855 @@
# for emacs: -*- mode: sh; -*-
# Armadillo May 2005 Broad Assembly
#
# DOWNLOAD SEQUENCE
ssh kkstore01
mkdir /cluster/store9/dasNov1
cd /cluster/data
ln -s /cluster/store9/dasNov1 dasNov1
cd /cluster/data/dasNov1
mkdir jkStuff bed
mkdir broad
cd broad
# ftp'ed with password from Broad
# -rw-rw-r-- 1 braney protein 69920541 May 20 07:52 assembly.agp
# -rw-rw-r-- 1 braney protein 745077193 May 20 08:15 assembly.bases.gz
# -rw-rw-r-- 1 braney protein 1222132418 May 20 09:08 assembly.quals.gz
# -rw-rw-r-- 1 braney protein 423586027 May 20 09:52 unplaced.fasta.gz
# -rw-rw-r-- 1 braney protein 1294098192 May 20 10:38 unplaced.qual.gz
gunzip assembly.bases.gz
faSize assembly.bases
# 2146362222 bases (0 N's 2146362222 real 2146362222 upper 0 lower) in 848432 sequences in 1 files
# Total size: mean 2529.8 sd 1972.8 min 61 (contig_80213) max 194887 (contig_17) median 2053
# N count: mean 0.0 sd 0.0
# U count: mean 2529.8 sd 1972.8
# L count: mean 0.0 sd 0.0
ssh kolossus
cd /cluster/data/dasNov1/broad
/cluster/bin/x86_64/agpAllToFaFile assembly.agp assembly.bases ../dasNov1.fa
cd ..
faSize dasNov1.fa
# 3856777364 bases (1710415142 N's 2146362222 real 2146362222 upper 0 lower) in 304391 sequences in 1 files
# Total size: mean 12670.5 sd 24184.9 min 1000 (scaffold_304389) max 974481 (scaffold_0) median 4960
# N count: mean 5619.1 sd 15102.0
# U count: mean 7051.3 sd 10915.0
# L count: mean 0.0 sd 0.0
/cluster/bin/scripts/agpToLift < assembly.agp > ../jkStuff/assembly.lft
# PARTITION SCAFFOLDS FOR REPEATMASKER RUN
# glom the tiny scaffolds up into ~50k collections
ssh kkstore01
cd /cluster/data/dasNov1
mkdir chunks50k
zcat broad/assembly.bases.gz | faSplit about stdin 50000 chunks50k/chunk_
cd chunks50k
for i in 0 1 2 3 4 5 6 7 8 9; do mkdir $i; mv *$i.fa $i; done
# RUN REPEAT MASKER
# make the run directory, output directory, and job list
ssh kkstore01
cd /cluster/data/dasNov1
cat << '_EOF_' > jkStuff/RMArmadillo
#!/bin/csh -fe
cd $1
/bin/mkdir -p /tmp/dasNov1/$2
/bin/cp $3 /tmp/dasNov1/$2/
pushd /tmp/dasNov1/$2
/cluster/bluearc/RepeatMasker/RepeatMasker -s -species eutheria $2
popd
/bin/cp /tmp/dasNov1/$2/$2.out ./
/bin/rm -fr /tmp/dasNov1/$2/*
/bin/rmdir --ignore-fail-on-non-empty /tmp/dasNov1/$2
/bin/rmdir --ignore-fail-on-non-empty /tmp/dasNov1
'_EOF_'
# << this line makes emacs coloring happy
chmod +x jkStuff/RMArmadillo
# mkdir RMRun RMOut
# for i in chunks800k/*.fa
# do
# d="/cluster/data/dasNov1"
# c=`basename $i`
# echo "../jkStuff/RMArmadillo $d/RMOut $c { check in line+ $d/$i} {check out line+ $d/RMOut/$c.out}"
# done > RMRun/RMJobs
mkdir RMRun1 RMOut1
cd RMOut1
mkdir 0 1 2 3 4 5 6 7 8 9
cd ../chunks50k
for i in */*.fa
do
e=`dirname $i`
d="/cluster/data/dasNov1"
c=`basename $i`
echo "../jkStuff/RMArmadillo $d/RMOut1/$e $c {check in line+ $d/chunks50k/$i} {check out line+ $d/RMOut1/$e/$c.out}"
done > ../RMRun1/RMJobs
# do the run
ssh kk9
cd /cluster/data/dasNov1/RMRun
para create RMJobs
para try, check, push, check,...
# Completed: 41361 of 41361 jobs
# CPU time in finished jobs: 52130147s 868835.79m 14480.60h 603.36d 1.653 y
# IO & Wait Time: 224553s 3742.55m 62.38h 2.60d 0.007 y
# Average job time: 1266s 21.10m 0.35h 0.01d
# Longest finished job: 6884s 114.73m 1.91h 0.08d
# Submission to last job: 150113s 2501.88m 41.70h 1.74d
# Lift up the split-scaffold .out's to scaffold .out's
ssh kkstore01
cd /cluster/data/dasNov1/RMOut1
for i in 0 1 2 3 4 5 6 7 8 9; do echo $i; liftUp -nohead $i.out ../jkStuff/assembly.lft warn $i/*.fa.out>/dev/null; done
head -3 0.out > dasNov1.out
for i in 0 1 2 3 4 5 6 7 8 9; do tail +4 $i.out; done >> dasNov1.out
# Load the .out files into the database with:
ssh hgwdev
hgLoadOut dasNov1 /cluster/data/dasNov1/RMOut1/dasNov1.out
# Loading up table dasNov1_rmsk
# note: 24 records dropped due to repStart > repEnd
hgsql dasNov1 -e 'rename table dasNov1_rmsk to rmsk'
# Fix up the indices too:
hgsql dasNov1 -e 'drop index bin on rmsk; \
drop index genoStart on rmsk; \
drop index genoEnd on rmsk; \
create index bin on rmsk (genoName(16), bin); \
create index genoStart on rmsk (genoName(16), genoStart); \
create index genoEnd on rmsk (genoName(16), genoEnd);'
# EXTRACTING GAP INFO FROM BLOCKS OF NS (DONE 11/5/04 angie)
ssh kkstore01
mkdir /cluster/data/dasNov1/bed/fakeAgp
cd /cluster/data/dasNov1/bed/fakeAgp
faGapSizes ../../downloads/scaffolds.fasta \
-niceSizes=5,10,20,25,30,40,50,100,250,500,1000,10000,100000
# Wow, all block of N's seem to be exactly 100bp long.
# hgFakeAgp's default -minContigGap of 25 will be fine.
hgFakeAgp ../../downloads/scaffolds.fasta fake.agp
ssh hgwdev
hgLoadGap -unsplit dasNov1 /cluster/data/dasNov1/bed/fakeAgp/fake.agp
# SIMPLE REPEATS (TRF)
ssh kkstore01
mkdir /cluster/data/dasNov1/bed/simpleRepeat
cd /cluster/data/dasNov1/bed/simpleRepeat
nice trfBig -trf=/cluster/bin/i386/trf ../../dasNov1.fa /dev/null -bedAt=simpleRepeat.bed -tempDir=/tmp \
| egrep -v '^(Removed|Tandem|Copyright|Loading|Allocating|Initializing|Computing|Scanning|Freeing)' > trf.log &
# check on this with
tail -f trf.log
# Load this into the database as so
ssh hgwdev
hgLoadBed dasNov1 simpleRepeat /cluster/data/dasNov1/bed/simpleRepeat/simpleRepeat.bed \
-sqlTable=$HOME/kent/src/hg/lib/simpleRepeat.sql
# FILTER SIMPLE REPEATS (TRF) INTO MASK
# make a filtered version of the trf output:
# keep trf's with period <= 12:
ssh kkstore01
cd /cluster/data/dasNov1/bed/simpleRepeat
awk '{if ($5 <= 12) print;}' simpleRepeat.bed > trfMask.bed
# MASK FA USING REPEATMASKER AND FILTERED TRF FILES
ssh kolossus
cd /cluster/data/dasNov1
nice /cluster/bin/x86_64/maskOutFa -soft dasNov1.fa bed/simpleRepeat/trfMask.bed dasNov1.simple.fa
nice /cluster/bin/x86_64/maskOutFa -softAdd dasNov1.simple.fa RMOut1/dasNov1.out dasNov1.masked.fa
# WARNING: negative rEnd: -104 scaffold_8757:558-754 L1M4
# WARNING: negative rEnd: -511 scaffold_34788:4357-4433 L1MCc
# WARNING: negative rEnd: -154 scaffold_34788:5061-5302 L1MCc
# WARNING: negative rEnd: -383 scaffold_119083:539-652 L1M4b
# WARNING: negative rEnd: -99 scaffold_181954:1627-2036 L1MCc
# WARNING: negative rEnd: -1033 scaffold_269:42483-42807 L1MEc
# WARNING: negative rEnd: -15 scaffold_18137:4036-4256 L1M4
# WARNING: negative rEnd: -301 scaffold_79493:4217-4508 L1MCc
# WARNING: negative rEnd: -542 scaffold_1800:103399-103624 L1M4b
# WARNING: negative rEnd: -594 scaffold_10038:2784-3058 L1MCc
# WARNING: negative rEnd: -916 scaffold_58908:40949-41369 L1MEc
# WARNING: negative rEnd: -247 scaffold_58908:41562-41685 L1MEc
# WARNING: negative rEnd: -99 scaffold_287090:3-415 L1MCc
# WARNING: negative rEnd: -162 scaffold_25:281259-281527 L1MCc
# WARNING: negative rEnd: -99 scaffold_45825:10398-10807 L1MCc
# WARNING: negative rEnd: -34 scaffold_47500:8156-8385 L1M4b
# WARNING: negative rEnd: -99 scaffold_67:113583-113881 L1MCc
mv dasNov1.fa dasNov1.unmasked.fa
# Now clean up the unmasked split scaffolds to avoid confusion later.
rm -r chunks500k scaffoldsSplit.fa jkStuff/scaffoldsSplit.lft
# CREATING DATABASE
# Create the database.
ssh hgwdev
# Make sure there is at least 5 gig free for the database
df -h /var/lib/mysql
Filesystem Size Used Avail Use% Mounted on
# /dev/sdc1 1.8T 915G 746G 56% /var/lib/mysql
hgsql '' -e 'create database dasNov1'
# STORE SEQUENCE AND ASSEMBLY INFORMATION
# Translate to 2bit
ssh kolossus
cd /cluster/data/dasNov1
/cluster/bin/x86_64/faToTwoBit dasNov1.masked.fa dasNov1.2bit
# Make chromInfo.tab.
mkdir bed/chromInfo
twoBitInfo dasNov1.2bit stdout | awk '{printf "%s\t%s\t/gbdb/dasNov1/dasNov1.2bit\n",$1,$2;}' \
| sort > bed/chromInfo/chromInfo.tab
# Make symbolic a link from /gbdb/dasNov1 to the 2bit.
ssh hgwdev
mkdir -p /gbdb/dasNov1
ln -s /cluster/data/dasNov1/dasNov1.2bit /gbdb/dasNov1/
# Load chromInfo table.
hgsql dasNov1 < $HOME/kent/src/hg/lib/chromInfo.sql
hgsql dasNov1 -e 'load data local infile "/cluster/data/dasNov1/bed/chromInfo/chromInfo.tab" into table chromInfo'
# Make chrom.sizes from chromInfo contents and check scaffold count.
hgsql dasNov1 -N -e 'select chrom,size from chromInfo' > /cluster/data/dasNov1/chrom.sizes
wc -l /cluster/data/dasNov1/chrom.sizes
# 304391 /cluster/data/dasNov1/chrom.sizes
# CREATING GRP TABLE FOR TRACK GROUPING
# Copy all the data from the table "grp"
# in an existing database to the new database
ssh hgwdev
hgsql dasNov1 -e 'create table grp (PRIMARY KEY(NAME)) select * from hg17.grp'
# MAKE HGCENTRALTEST ENTRY AND TRACKDB TABLE
# Warning: genome and organism fields must correspond
# with defaultDb values
echo 'INSERT INTO genomeClade (genome,clade,priority) values ("Armadillo", "vertebrate",56);' \
| hgsql -h genome-testdb hgcentraltest
echo 'INSERT INTO dbDb (name, description, nibPath, organism, defaultPos, active, orderKey, genome, scientificName, \
htmlPath, hgNearOk, hgPbOk, sourceName) values ("dasNov1", "May. 2005", "/gbdb/dasNov1", "Armadillo", \
"", 1, 57, "Armadillo", "Dasypus novemcinctus", "/gbdb/dasNov1/html/description.html", 0, 0, "Broad May 2005");'\
| hgsql -h genome-testdb hgcentraltest
echo 'INSERT INTO defaultDb (genome, name) values ("Armadillo", "dasNov1");' \
| hgsql -h genome-testdb hgcentraltest
# Make trackDb table so browser knows what tracks to expect:
ssh hgwdev
cd ~/kent/src/hg/makeDb/trackDb
cvs up -d -P
# Edit trackDb/makefile to add dasNov1 to the DBS variable.
mkdir drosophila/dasNov1
# Create a simple drosophila/dasNov1/description.html file.
cvs add drosophila/dasNov1
cvs add drosophila/dasNov1/description.html
make update DBS=dasNov1 ZOO_DBS=
# go public on genome-test
cvs ci makefile
cvs ci drosophila/dasNov1
mkdir /gbdb/dasNov1/html
# in a clean, updated tree's kent/src/hg/makeDb/trackDb:
make alpha
# PUT SEQUENCE ON /ISCRATCH FOR BLASTZ (DONE 11/3/04 angie)
# First, agglomerate small scaffolds into chunks of ~100k median
# (many scaffolds are larger than that) so we don't have too many
# files for one dir, but keep a reasonably low job run time:
ssh kkstore01
cd /cluster/data/dasNov1
mkdir chunksUnsplit
faSplit about scaffolds.fa 100000 chunksUnsplit/chunk_
ssh kkr1u00
mkdir /iscratch/i/dasNov1
cp -pR /cluster/data/dasNov1/chunksUnsplit /iscratch/i/dasNov1/
cp -p /cluster/data/dasNov1/dasNov1.2bit /iscratch/i/dasNov1/
iSync
# PRODUCING GENSCAN PREDICTIONS (DONE 11/4/04 angie)
ssh kkstore01
# Make hard-masked scaffolds and split up for processing:
cd /cluster/data/dasNov1
maskOutFa scaffolds.fa hard scaffolds.fa.masked
mkdir chunksUnsplitMasked
faSplit about scaffolds.fa.masked 100000 chunksUnsplitMasked/chunk_
mkdir /cluster/data/dasNov1/bed/genscan
cd /cluster/data/dasNov1/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 ../../chunksUnsplitMasked/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/dasNov1/bed/genscan
para create jobList
para try, check, push, check, ...
#Completed: 463 of 463 jobs
#Average job time: 12s 0.21m 0.00h 0.00d
#Longest job: 317s 5.28m 0.09h 0.00d
#Submission to last job: 445s 7.42m 0.12h 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 kkstore01
cd /cluster/data/dasNov1/bed/genscan
cat gtf/*.gtf > genscan.gtf
cat subopt/*.bed > genscanSubopt.bed
cat pep/*.pep > genscan.pep
# Clean up
rm -r /cluster/data/dasNov1/chunksUnsplitMasked
# Load into the database as so:
ssh hgwdev
cd /cluster/data/dasNov1/bed/genscan
# Reloaded without -genePredExt 1/6/05:
ldHgGene -gtf dasNov1 genscan genscan.gtf
hgPepPred dasNov1 generic genscanPep genscan.pep
hgLoadBed dasNov1 genscanSubopt genscanSubopt.bed
# MAKE DOWNLOADABLE FILES (DONE 11/4/04 angie)
# RECREATE BIGZIPS DOWNLOADS AND README FILE (DONE, 2006-05-05, hartera)
# ADDED md5sum.txt FILE FOR LIFTOVER DOWNLOADS AND CREATED CORRECT md5sum.txt
# FOR vsMm7 AND vsHg18 DOWNLOADS (DONE, 2006-05-23, hartera)
ssh kkstore01
mkdir /cluster/data/dasNov1/zips
cd /cluster/data/dasNov1
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/dasNov1
cd /usr/local/apache/htdocs/goldenPath/dasNov1
mkdir bigZips database
# Create README.txt files in bigZips/ and database/ to explain the files.
cd bigZips
cp -p /cluster/data/dasNov1/zips/*.zip .
md5sum *.zip > md5sum.txt
# Add more bigZips downloads. Some of the above downloads don't exist
# anymore in bigZips in /usr/local/apache/htdocs/goldenPath/... on hgwdev
# (2006-05-05, hartera)
ssh kkstore04
mkdir /cluster/data/dasNov1/bigZips
cd /cluster/data/dasNov1
# soft-masked scaffolds sequences
cp -p dasNov1.masked.fa.gz ./bigZips/scaffoldSoftMask.fa.gz
# assembly agp
cp -p ./broad/assembly.agp.gz ./bigZips/
# Simple Repeats
cp -p ./bed/simpleRepeat/trfMask.bed ./bigZips/trf.bed
# RepeatMasker output
cp -p ./RMOut1/dasNov1.out ./bigZips/rmsk.out
# unmasked scaffolds sequences
cp -p dasNov1.unmasked.fa.gz ./bigZips/scaffold.fa.gz
cd bigZips
gzip trf.bed
gzip rmsk.out
# check integrity of files
foreach f (*.gz)
echo $f
gunzip -t $f | tail -1
end
md5sum *.gz > md5sum.txt
# link the *.gz and *.txt files to hgwdev:/usr/local/apache/....
ssh hgwdev
set gp=/usr/local/apache/htdocs/goldenPath/dasNov1
rm -r $gp/bigZips
mkdir -p $gp/bigZips
ln -s /cluster/data/dasNov1/bigZips/{*.gz,*.txt} $gp/bigZips
# copy over README.txt and edit for bigZips
# dasNov1ToMm7.over.chain.gz is in the md5sum.txt in the
# vsMm7 directory. It should be in the liftOver directory so move it there
# and re-make md5sum.txt for vsMm7 dir.
# Add liftOver md5sum.txt (hartera, 2006-05-23)
ssh hgwdev
set gp=/usr/local/apache/htdocs/goldenPath/dasNov1
# copy over from hgwdevold
cd /cluster/data/dasNov1/bed/liftOver
scp hgwdevold:$gp/liftOver/dasNov1ToMm7.over.chain.gz .
ln -s /cluster/data/dasNov1/bed/liftOver/*.gz $gp/liftOver
cd $gp/liftOver
# create md5sum.txt
md5sum *.gz > md5sum.txt
# recreate md5sum.txt for vsMm7 files
cd $gp/vsMm7
rm md5sum.txt
md5sum *.gz > md5sum.txt
# recreate md5sum.txt for vsHg18 files as this includes files in an
# axtNet directory that is not in this directory.
cd $gp/vsHg18
rm md5sum.txt
md5sum *.gz > md5sum.txt
# Check that README.txt is correct for vsMm7 and vsHg18.
# SWAP DM1-DROANA1 BLASTZ (DONE 11/4/04 angie)
ssh kkstore01
mkdir /cluster/data/dasNov1/bed/blastz.dm1.swap.2004-11-03
ln -s blastz.dm1.swap.2004-11-03 /cluster/data/dasNov1/bed/blastz.dm1
cd /cluster/data/dasNov1/bed/blastz.dm1
set aliDir = /cluster/data/dm1/bed/blastz.dasNov1
cp $aliDir/S1.len S2.len
cp $aliDir/S2.len S1.len
# With 11k 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.
cat $aliDir/axtChrom/chr*.axt \
| axtSwap stdin $aliDir/S1.len $aliDir/S2.len stdout \
| axtSort stdin dm1.axt
du -sh $aliDir/axtChrom dm1.axt
#389M /cluster/data/dm1/bed/blastz.dasNov1/axtChrom
#389M dm1.axt
# CHAIN MELANOGASTER BLASTZ (DONE 11/4/04 angie)
# Run axtChain on kolossus (one big dm1.axt input)
ssh kolossus
mkdir /cluster/data/dasNov1/bed/blastz.dm1/axtChain
cd /cluster/data/dasNov1/bed/blastz.dm1/axtChain
axtChain -verbose=0 ../dm1.axt /cluster/data/dasNov1/dasNov1.2bit \
/cluster/data/dm1/nib stdout \
| chainAntiRepeat /cluster/data/dasNov1/dasNov1.2bit \
/cluster/data/dm1/nib stdin stdout \
| chainMergeSort stdin > all.chain
# Load chains into database
ssh hgwdev
cd /cluster/data/dasNov1/bed/blastz.dm1/axtChain
hgLoadChain -tIndex dasNov1 chainDm1 all.chain
# NET MELANOGASTER BLASTZ (DONE 11/4/04 angie)
ssh kkstore01
cd /cluster/data/dasNov1/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/dasNov1/bed/blastz.dm1/axtChain
netClass -noAr noClass.net dasNov1 dm1 melanogaster.net \
|& g -v "table gap doesn't exist"
# Make a 'syntenic' subset:
ssh kkstore01
cd /cluster/data/dasNov1/bed/blastz.dm1/axtChain
rm noClass.net
netFilter -syn melanogaster.net > melanogasterSyn.net
# Load the nets into database
ssh hgwdev
cd /cluster/data/dasNov1/bed/blastz.dm1/axtChain
netFilter -minGap=10 melanogaster.net | hgLoadNet dasNov1 netDm1 stdin
netFilter -minGap=10 melanogasterSyn.net \
| hgLoadNet dasNov1 netSyntenyDm1 stdin
# MAKE AXTNET (DONE 11/4/04 angie)
ssh kkstore01
cd /cluster/data/dasNov1/bed/blastz.dm1/axtChain
netToAxt melanogaster.net all.chain /cluster/data/dasNov1/dasNov1.2bit \
/cluster/data/dm1/nib stdout \
| axtSort stdin melanogasterNet.axt
# MAKE VSDM1 DOWNLOADABLES (DONE 11/4/04 angie)
ssh kkstore01
cd /cluster/data/dasNov1/bed/blastz.dm1/axtChain
nice gzip *.{chain,net,axt}
ssh hgwdev
mkdir /usr/local/apache/htdocs/goldenPath/dasNov1/vsDm1
cd /usr/local/apache/htdocs/goldenPath/dasNov1/vsDm1
cp -p /cluster/data/dasNov1/bed/blastz.dm1/axtChain/all.chain.gz \
melanogaster.chain.gz
cp -p /cluster/data/dasNov1/bed/blastz.dm1/axtChain/melanogaster.net.gz .
cp -p /cluster/data/dasNov1/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 11/4/04 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/dasNov1
blat /cluster/data/dasNov1/dasNov1.2bit /dev/null /dev/null -tileSize=11 \
-makeOoc=/cluster/bluearc/dasNov1/11.ooc -repMatch=100
#Wrote 9721 overused 11-mers to /cluster/bluearc/dasNov1/11.ooc
cp -p /cluster/bluearc/dasNov1/*.ooc /iscratch/i/dasNov1/
iSync
# GET GENBANK mRNA AND EST COUNTS (DONE 11/8/04 angie)
# Go to the latest GenBank full release dir and get an idea of how
# many mRNAs and ESTs there are to align.
ssh eieio
cd /cluster/data/genbank/data/processed/genbank.144.0/full
awk '$4 == "Drosophila" {print $4 " " $5;}' mrna.gbidx | sort | uniq -c
# 9 Drosophila ananassae
# 1 Drosophila mojavensis
# 33 Drosophila virilis
# Wow, questionable whether we should have a native mRNA track here.
awk '$4 == "Drosophila" {print $4 " " $5;}' est*.gbidx | sort | uniq -c
# 382439 Drosophila melanogaster
# 4105 Drosophila simulans
# 779 Drosophila yakuba
# And a native EST track isn't even a possibility for the new flies
# at this point!
# AUTO UPDATE GENBANK MRNA RUN (DONE 11/16/04 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):
# dasNov1 (D. ananassae)
dasNov1.genome = /iscratch/i/dasNov1/dasNov1.2bit
dasNov1.mondoTwoBitParts = 1000
dasNov1.lift = no
dasNov1.refseq.mrna.native.load = no
dasNov1.refseq.mrna.xeno.load = yes
dasNov1.refseq.mrna.xeno.pslReps = -minCover=0.15 -minAli=0.75 -nearTop=0.005
dasNov1.genbank.mrna.xeno.load = yes
# GenBank has no D. ananassae ESTs at this point... that may change.
dasNov1.genbank.est.native.load = no
dasNov1.genbank.est.xeno.load = no
dasNov1.downloadDir = dasNov1
dasNov1.perChromTables = no
cvs ci etc/genbank.conf
# Since D. ananassae 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/dasNov1/11.ooc
cvs diff src/align/gbBlat
make
cvs ci src/align/gbBlat
# Install to /cluster/data/genbank:
make install-server
ssh `fileServer /cluster/data/genbank/`
cd /cluster/data/genbank
# This is an -initial run, (xeno) RefSeq only:
nice bin/gbAlignStep -srcDb=refseq -type=mrna -initial dasNov1 &
tail -f [its logfile]
# Load results:
ssh hgwdev
cd /cluster/data/genbank
nice bin/gbDbLoadStep -verbose=1 -drop -initialLoad dasNov1
featureBits dasNov1 xenoRefGene
#16520385 bases of 165766797 (9.966%) in intersection
# Clean up:
rm -rf work/initial.dasNov1
# This is an -initial run, mRNA only:
nice bin/gbAlignStep -srcDb=genbank -type=mrna -initial dasNov1 &
tail -f [its logfile]
# Load results:
ssh hgwdev
cd /cluster/data/genbank
nice bin/gbDbLoadStep -verbose=1 -drop -initialLoad dasNov1
featureBits dasNov1 all_mrna
#19602 bases of 165766797 (0.012%) in intersection
featureBits dasNov1 xenoMrna
#17295487 bases of 165766797 (10.434%) in intersection
# Clean up:
rm -rf work/initial.dasNov1
# MAKE GCPERCENT (DONE 11/4/04 angie)
ssh hgwdev
mkdir /cluster/data/dasNov1/bed/gcPercent
cd /cluster/data/dasNov1/bed/gcPercent
# create and load gcPercent table
hgGcPercent dasNov1 /cluster/data/dasNov1
# MAKE HGCENTRALTEST BLATSERVERS ENTRY (DONE 7-20-05 braney)
ssh hgwdev
echo 'insert into blatServers values("dasNov1", "blat2", "17780", 1, 0); \
insert into blatServers values("dasNov1", "blat2", "17781", 0, 1);' \
| hgsql -h genome-testdb hgcentraltest
# MAKE Drosophila Proteins track (DONE braney 11/17/04)
ssh kkstore01
mkdir -p /cluster/data/dasNov1/blastDb
cd /cluster/data/dasNov1/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/dasNov1/blastDb
cp /cluster/data/dasNov1/blastDb/* /iscratch/i/dasNov1/blastDb
(iSync) 2>&1 > sync.out
mkdir -p /cluster/data/dasNov1/bed/tblastn.dm1FB
cd /cluster/data/dasNov1/bed/tblastn.dm1FB
ls -1S /iscratch/i/dasNov1/blastDb/*.nsq | sed "s/\.nsq//" > bug.lst
exit
# back to kkstore01
cd /cluster/data/dasNov1/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/dasNov1/bed/tblastn.dm1FB
para create blastSpec
para try, push
# Completed: 151668 of 151668 jobs
# CPU time in finished jobs: 2932565s 48876.08m 814.60h 33.94d 0.093 y
# IO & Wait Time: 694006s 11566.77m 192.78h 8.03d 0.022 y
# Average job time: 24s 0.40m 0.01h 0.00d
# Longest job: 2721s 45.35m 0.76h 0.03d
# Submission to last job: 73860s 1231.00m 20.52h 0.85d
cat << '_EOF_' > chainGsub
#LOOP
chainSome $(path1)
#ENDLOOP
'_EOF_'
cat << '_EOF_' > chainSome
(cd $1; cat q.*.psl | simpleChain -prot -outPsl -maxGap=25000 stdin ../c.`basename $1`.psl)
'_EOF_'
chmod +x chainSome
ls -1dS `pwd`/blastOut/fb?? > chain.lst
gensub2 chain.lst single chainGsub chainSpec
para create chainSpec
# should run this on the mini-cluster or with my shove script
# so you can limit the number of jobs starting to 3 or 4
para try, push...
# Completed: 383 of 383 jobs
# CPU time in finished jobs: 327s 5.44m 0.09h 0.00d 0.000 y
# IO & Wait Time: 8218s 136.97m 2.28h 0.10d 0.000 y
# Average job time: 22s 0.37m 0.01h 0.00d
# Longest job: 54s 0.90m 0.01h 0.00d
# Submission to last job: 674s 11.23m 0.19h 0.01d
exit
# back to kkstore01
cd /cluster/data/dasNov1/bed/tblastn.dm1FB/blastOut
for i in fb??
do
awk "(\$13 - \$12)/\$11 > 0.6 {print}" c.$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/dasNov1/bed/tblastn.dm1FB/blastDm1FB.psl
ssh hgwdev
cd /cluster/data/dasNov1/bed/tblastn.dm1FB
hgLoadPsl dasNov1 blastDm1FB.psl
# End tblastn
# SWAP CHAINS FROM DM2, BUILD NETS ETC. (DONE 3/2/05 angie)
ssh kkstore01
mkdir /cluster/data/dasNov1/bed/blastz.dm2.swap
cd /cluster/data/dasNov1/bed/blastz.dm2.swap
doBlastzChainNet.pl -swap /cluster/data/dm2/bed/blastz.dasNov1/DEF \
>& do.log &
tail -f do.log
# Add {chain,net}Dm2 to trackDb.ra if necessary.
# Add /usr/local/apache/htdocs/goldenPath/dasNov1/vsDm2/README.txt
# CHAIN AND NET SWAPPED HUMAN BLASTZ (WORKING 12/22/05 kate)
# Working in Brian's blastz dir (not doced).
# This procedure follows conventions in doBlastzNetChain.pl,
# so it can be used to complete the processing
ssh kki
cd /cluster/data/dasNov1/bed/zb.hg17
ln -s `pwd` /cluster/data/hg17/bed/blastz.dasNov1
mkdir -p axtChain/run/chain
ls -1S /cluster/data/dasNov1/bed/zb.hg17/axtChrom/*.axt.gz | \
sed 's/.gz//' > axtChain/run/input.lst
cd axtChain/run
cat > chain.csh << 'EOF'
#!/bin/csh -ef
zcat $1 | \
axtChain -verbose=0 -minScore=3000 -linearGap=medium stdin \
/cluster/data/hg17/nib /cluster/data/dasNov1/dasNov1.2bit stdout | \
chainAntiRepeat /cluster/data/hg17/nib /cluster/data/dasNov1/dasNov1.2bit \
stdin $2
'EOF'
# << happy emacs
cat > gsub << 'EOF'
#LOOP
chain.csh {check in exists $(path1).gz} {check out line+ chain/$(root1).chain}
#ENDLOOP
'EOF'
# << happy emacs
chmod a+x chain.csh
gensub2 input.lst single gsub jobList
para create jobList; para try
para check
para push
# wait for completion
para time > run.time
# finish pipeline
cd /cluster/data/dasNov1/bed/zb.hg17
/cluster/bin/scripts/doBlastzChainNet.pl -verbose=2 \
-continue=chainMerge \
-bigClusterHub=pk \
`pwd`/DEF >& blastz.out &
############################################################################
## BLASTZ swap from mm9 alignments (2007-11-11 - markd)
ssh hgwdev
mkdir /cluster/data/dasNov1/bed/blastz.mm9.swap
cd /cluster/data/dasNov1/bed/blastz.mm9.swap
ln -s blastz.mm9.swap ../blastz.mm9
/cluster/bin/scripts/doBlastzChainNet.pl \
-swap /cluster/data/mm9/bed/blastz.dasNov1/DEF >& swap.out&
nice +19 featureBits dasNov1 chainMm9Link
# 438318232 bases of 2146362222 (20.421%) in intersection
############################################################################
# TRANSMAP vertebrate.2008-05-20 build (2008-05-24 markd)
vertebrate-wide transMap alignments were built Tracks are created and loaded
by a single Makefile. This is available from:
svn+ssh://hgwdev.cse.ucsc.edu/projects/compbio/usr/markd/svn/projs/transMap/tags/vertebrate.2008-05-20
see doc/builds.txt for specific details.
############################################################################
############################################################################
# TRANSMAP vertebrate.2008-06-07 build (2008-06-30 markd)
vertebrate-wide transMap alignments were built Tracks are created and loaded
by a single Makefile. This is available from:
svn+ssh://hgwdev.cse.ucsc.edu/projects/compbio/usr/markd/svn/projs/transMap/tags/vertebrate.2008-06-30
see doc/builds.txt for specific details.
############################################################################
# dasNov1 - Armadillo - Ensembl Genes version 51 (DONE - 2008-12-03 - hiram)
ssh kkr14u07
cd /hive/data/genomes/dasNov1
cat << '_EOF_' > dasNov1.ensGene.ra
# required db variable
db dasNov1
# do we need to translate geneScaffold coordinates
geneScaffolds yes
# ignore genes that do not properly convert to a gene pred, and contig
# names that are not in the UCSC assembly
skipInvalid yes
# ignore the three genes that have invalid structures from Ensembl:
# 18192: ENSDNOT00000004471 no exonFrame on CDS exon 7
# 35247: ENSDNOT00000007696 no exonFrame on CDS exon 8
# 38952: ENSDNOT00000019234 no exonFrame on CDS exon 0
'_EOF_'
# << happy emacs
doEnsGeneUpdate.pl -ensVersion=51 dasNov1.ensGene.ra
ssh hgwdev
cd /hive/data/genomes/dasNov1/bed/ensGene.51
featureBits dasNov1 ensGene
# 22658142 bases of 2146362222 (1.056%) in intersection
*** All done! (through the 'makeDoc' step)
*** Steps were performed in /hive/data/genomes/dasNov1/bed/ensGene.51
############################################################################
+############################################################################
+# TRANSMAP vertebrate.2009-07-01 build (2009-07-21 markd)
+
+vertebrate-wide transMap alignments were built Tracks are created and loaded
+by a single Makefile. This is available from:
+ svn+ssh://hgwdev.cse.ucsc.edu/projects/compbio/usr/markd/svn/projs/transMap/tags/vertebrate.2009-07-01
+
+see doc/builds.txt for specific details.
+############################################################################