src/hg/makeDb/doc/aplCal1.txt 1.6
1.6 2009/06/09 00:45:27 braney
human proteins
Index: src/hg/makeDb/doc/aplCal1.txt
===================================================================
RCS file: /projects/compbio/cvsroot/kent/src/hg/makeDb/doc/aplCal1.txt,v
retrieving revision 1.5
retrieving revision 1.6
diff -b -B -U 1000000 -r1.5 -r1.6
--- src/hg/makeDb/doc/aplCal1.txt 28 May 2009 18:21:39 -0000 1.5
+++ src/hg/makeDb/doc/aplCal1.txt 9 Jun 2009 00:45:27 -0000 1.6
@@ -1,342 +1,506 @@
#########################################
#
# aplCal1 = Aplysia californica
#
# (Galt)
# set up main genome directory
ssh hgwdev
cd /hive/data/genomes
mkdir aplCal1
cd aplCal1
mkdir download
cd download
# get sequence from BROAD
# http://www.broad.mit.edu/ftp/pub/assemblies/invertebrates/aplysia/
cat << '_EOF_' > fetch.sh
#!/bin/sh
wget --timestamping -r -np -l 2 -nd -L
'ftp://ftp.broad.mit.edu/pub/assemblies/invertebrates/aplysia'
'_EOF_'
# << happy emacs
chmod +x fetch.sh
./fetch.sh
# fixup ids for super fasta and quals
# change ">scaffold_0.1-1784514 (Draft_v1)"
# to ">scaffold_0"
# sed 's/\(>scaffold_[0-9]*\)[.].*/\1/'
cat << '_EOF_' > fix.csh
#!/bin/tcsh
gunzip -c assembly_supers.fasta.gz | sed 's/\(>scaffold_[0-9]*\)[.].*/\1/' \
| gzip -c > assembly_supers.fasta.fix.gz
gunzip -c assembly_supers.qual.gz | sed 's/\(>scaffold_[0-9]*\)[.].*/\1/' \
| gzip -c > assembly_supers.qual.fix.gz
'_EOF_'
# << happy emacs
chmod +x fix.csh
./fix.csh
# totalKinds is something I made to add up the numbers of each kind
# it just reveals that there are too many unplaced fragments
# to consider making them all scaffolds, and as a chrUn it would
# be too large. We are going to just leave these out of the
# assembly like Broad did. JK says ok.
totalKinds assembly.unplaced 2 > unplaced.types
cat unplaced.types
#unplaced 4270994
#low_quality 945330
#other 1781
#deliberate 273497
#vector_or_host 63932
#multiple_instance 27153
# back to the main directory
cd /hive/data/genomes/aplCal1
# Run automation to make the basic genome
cat << '_EOF_' > aplCal1.config.ra
# Config parameters for makeGenomeDb.pl:
db aplCal1
scientificName Aplysia californica
commonName Sea Hare
assemblyDate Sept. 2008
assemblyLabel Broad Institute v. 1.0
orderKey 825
clade other
genomeCladePriority 19
mitoAcc NC_005827
fastaFiles /hive/data/genomes/aplCal1/download/assembly_supers.fasta.fix.gz
agpFiles /hive/data/genomes/aplCal1/download/assembly.agp
qualFiles /hive/data/genomes/aplCal1/download/assembly_supers.qual.fix.gz
dbDbSpeciesDir seaHare
taxId 6500
'_EOF_'
# << happy emacs
time makeGenomeDb.pl aplCal1.config.ra > & makeGenomeDb.pl.out &
# took less than 10 minutes
featureBits -countGaps aplCal1 gap
#96577943 bases of 715806041 (13.492%) in intersection
cat chrom.sizes | awk '{sum+=$2;print sum,$0}'
#715806041 scaffold_8765 5001
# same total
# Organism Image
wget -O /usr/local/apache/htdocs/images/Aplysia_californica.jpg \
'http://upload.wikimedia.org/wikipedia/commons/thumb/e/ef/Aplysia_californica.jpg/250px-Aplysia_californica.jpg'
# Edit and check-in templates for description.html, gold.html, gap.html, aplCal1/trackDb.ra
# repeat mask
time doRepeatMasker.pl aplCal1 > & doRepeatMasker.pl.out &
cat bed/RepeatMasker.2009-05-13/faSize.rmsk.txt
#715806041 bases (96577949 N's 619228092 real 559834485 upper 59393607 lower)
#in 8767 sequences in 1 files
[hgwdev:aplCal1> featureBits -countGaps aplCal1 rmsk
59403223 bases of 715806041 (8.299%) in intersection
# simple repeat masker trf
time doSimpleRepeat.pl aplCal1 > & doSimpleRepeat.pl.out &
[hgwdev:aplCal1> featureBits -countGaps aplCal1 simpleRepeat
#55409917 bases of 715806041 (7.741%) in intersection
# make final masked .2bit
[hgwdev:aplCal1> twoBitMask aplCal1.rmsk.2bit -add bed/simpleRepeat.2009-05-13/trfMask.bed aplCal1.2bit
#Warning: BED file bed/simpleRepeat.2009-05-13/trfMask.bed has >=13 fields
#which means it might contain block coordinates, but this program uses only the
#first three fields (the entire span -- no support for blocks).
############################################################################
# prepare cluster data (DONE - 2009-05-19 - Galt)
ssh hgwdev
cd /hive/data/genomes/aplCal1
# create gbdb symlink
rm /gbdb/aplCal1/aplCal1.2bit
ln -s `pwd`/aplCal1.2bit /gbdb/aplCal1/
time blat aplCal1.2bit \
/dev/null /dev/null -tileSize=11 -makeOoc=11.ooc -repMatch=1024
#Wrote 2216 overused 11-mers to 11.ooc
#29.646u 0.750s 0:33.12 91.7% 0+0k 0+0io 3pf+0w
mkdir /hive/data/staging/data/aplCal1
cp -p aplCal1.2bit /hive/data/staging/data/aplCal1
cp -p 11.ooc /hive/data/staging/data/aplCal1
cp -p chrom.sizes /hive/data/staging/data/aplCal1
# ask admin to sync this directory: /hive/data/staging/data/aplCal1/
# to the kluster nodes /scratch/data/aplCal1/
############################################################################
# running cpgIsland (DONE - 2009-05-19 - Galt)
ssh hgwdev
cd /hive/data/genomes/aplCal1
mkdir bed/cpgIsland
cd bed/cpgIsland
cvs -d /projects/compbio/cvsroot checkout -P hg3rdParty/cpgIslands
cd hg3rdParty/cpgIslands
# comment out the following two lines if it compiles cleanly
# some day (there were some other fixups too, adding include lines)
sed -i -e "s#\(extern char\* malloc\)#// \1#" cpg_lh.c
make
#warning: incompatible implicit declaration of built-in function
# ignore the warnings
cd ../../
ln -s hg3rdParty/cpgIslands/cpglh.exe
# make hardmasked fasta files
mkdir -p hardMaskedFa
bash
cut -f1 ../../chrom.sizes | while read C
do
echo ${C}
twoBitToFa ../../aplCal1.2bit:$C stdout \
| maskOutFa stdin hard hardMaskedFa/${C}.fa
done
#exit bash
exit
cut -f1 ../../chrom.sizes > chr.list
cat << '_EOF_' > template
#LOOP
./runOne $(root1) {check out line results/$(root1).cpg}
#ENDLOOP
'_EOF_'
# << happy emacs
cat << '_EOF_' > runOne
#!/bin/csh -fe
./cpglh.exe hardMaskedFa/$1.fa > /scratch/tmp/$1.$$
mv /scratch/tmp/$1.$$ $2
'_EOF_'
# << happy emacs
mkdir results
chmod +x runOne
pk
cd /hive/data/genomes/aplCal1/bed/cpgIsland
gensub2 chr.list single template jobList
para create jobList
para try
para check
para push
para time
#Completed: 8767 of 8767 jobs
#CPU time in finished jobs: 47s 0.79m 0.01h 0.00d 0.000 y
#IO & Wait Time: 43996s 733.26m 12.22h 0.51d 0.001 y
#Average job time: 5s 0.08m 0.00h 0.00d
#Longest finished job: 26s 0.43m 0.01h 0.00d
#Submission to last job: 1161s 19.35m 0.32h 0.01d
# Transform cpglh output to bed +
catDir results | awk '{ \
$2 = $2 - 1; \
width = $3 - $2; \
printf("%s\t%d\t%s\t%s %s\t%s\t%s\t%0.0f\t%0.1f\t%s\t%s\n", \
$1, $2, $3, $5,$6, width, \
$6, width*$7*0.01, 100.0*2*$6/width, $7, $9); \
}' > cpgIsland.bed
# took around 15 minutes
cd /hive/data/genomes/aplCal1/bed/cpgIsland
hgLoadBed aplCal1 cpgIslandExt -tab \
-sqlTable=$HOME/kent/src/hg/lib/cpgIslandExt.sql cpgIsland.bed
#Reading cpgIsland.bed
#Loaded 23247 elements of size 10
#Sorted
#Creating table definition for cpgIslandExt
#Saving bed.tab
#Loading aplCal1
#############################################################
#
# Since there is no chrUn and no huge gaps,
# MarkD and Hiram say we can skip the unbridged-lift step
#
############################################################################
# AUTO UPDATE GENBANK RUN (DONE - 2009-05-27,13 - Galt)
# align with latest genbank process.
# first time species step
cd ~/kent/src/hg/makeDb/genbank/src/lib
vi gbGenome.c
# add aplCal
cd ~/kent/src/hg/makeDb/genbank
make server-install
cd ~/kent/src/hg/makeDb/genbank
cvsup
# edit etc/genbank.conf to add aplCal1
# aplCal1 - (Sea Hare Broad WGS Aplcal 2.0)
# Assembly Accession: AASC00000000.2
aplCal1.serverGenome = /hive/data/genomes/aplCal1/aplCal1.2bit
aplCal1.clusterGenome = /scratch/data/aplCal1/aplCal1.2bit
aplCal1.ooc = /scratch/data/aplCal1/11.ooc
aplCal1.refseq.mrna.native.pslCDnaFilter = ${lowCover.refseq.mrna.native.pslCDnaFilter}
aplCal1.refseq.mrna.xeno.pslCDnaFilter = ${lowCover.refseq.mrna.xeno.pslCDnaFilter}
aplCal1.genbank.mrna.native.pslCDnaFilter = ${lowCover.genbank.mrna.native.pslCDnaFilter}
aplCal1.genbank.mrna.xeno.pslCDnaFilter = ${lowCover.genbank.mrna.xeno.pslCDnaFilter}
aplCal1.genbank.est.native.pslCDnaFilter = ${lowCover.genbank.est.native.pslCDnaFilter}
aplCal1.genbank.est.xeno.pslCDnaFilter = ${lowCover.genbank.est.xeno.pslCDnaFilter}
aplCal1.genbank.est.native.load = yes
aplCal1.genbank.est.xeno.load = no
aplCal1.refseq.mrna.xeno.load = yes
aplCal1.refseq.mrna.xeno.loadDesc = yes
aplCal1.perChromTables = no
aplCal1.downloadDir = aplCal1
aplCal1.genbank.mrna.blatTargetDb = yes
aplCal1.lift = no
cvs ci -m "Added aplCal1." etc/genbank.conf
# update /cluster/data/genbank/:
make etc-update
ssh genbank
screen # use a screen to manage this job
cd /cluster/data/genbank
time nice +19 bin/gbAlignStep -initial aplCal1 &
# real 3:21:30.93
# var/build/logs/2009.05.27-20:11:50.aplCal1.initalign.log
# load database when finished
ssh hgwdev
screen # use screen to manage this long running command
cd /cluster/data/genbank
time nice +19 ./bin/gbDbLoadStep -drop -initialLoad aplCal1 &
# logFile: var/dbload/hgwdev/logs/2009.05.27-23:42:23.dbload.log
# real 23:26.57
# enable daily alignment and update of hgwdev
cd ~/kent/src/hg/makeDb/genbank
cvsup
# add aplCal1 to:
etc/align.dbs
etc/hgwdev.dbs
cvs ci -m "Added aplCal1" etc/align.dbs etc/hgwdev.dbs
make etc-update
#########################################################################
# BLATSERVERS ENTRY (DONE - 2009-05-28 - Galt)
# After getting a blat server assigned by the Blat Server Gods,
ssh hgwdev
hgsql -e 'INSERT INTO blatServers (db, host, port, isTrans, canPcr) \
VALUES ("aplCal1", "blat13", "17822", "1", "0"); \
INSERT INTO blatServers (db, host, port, isTrans, canPcr) \
VALUES ("aplCal1", "blat13", "17823", "0", "1");' \
hgcentraltest
# test it with some sequence
############################################################################
# Making download files (DONE - 2009-05-28 - Galt)
cd /hive/data/genomes/aplCal1
ln -s bed/RepeatMasker.2009-05-13/aplCal1.fa.out .
cd bed
ln -s simpleRepeat.2009-05-13 simpleRepeat
cd ..
makeDownloads.pl aplCal1 >& downloads.log
(*** book mark of stuff actually done to here ***)
+###########################################################################
+# HUMAN (hg18) PROTEINS TRACK (DONE 2009-06-08 braney )
+ # bash if not using bash shell already
+
+ cd /cluster/data/aplCal1
+ mkdir /cluster/data/aplCal1/blastDb
+
+ awk '{if ($2 > 1000000) print $1}' chrom.sizes > 1meg.lst
+ twoBitToFa -seqList=1meg.lst aplCal1.2bit temp.fa
+ faSplit gap temp.fa 1000000 blastDb/x -lift=blastDb.lft
+ rm temp.fa 1meg.lst
+
+ awk '{if ($2 <= 1000000) print $1}' chrom.sizes > less1meg.lst
+ twoBitToFa -seqList=less1meg.lst aplCal1.2bit temp.fa
+ faSplit about temp.fa 1000000 blastDb/y
+ rm temp.fa less1meg.lst
+
+ cd blastDb
+ for i in *.fa
+ do
+ /hive/data/outside/blast229/formatdb -i $i -p F
+ done
+ rm *.fa
+ ls *.nsq | wc -l
+# 656
+
+
+ mkdir -p /cluster/data/aplCal1/bed/tblastn.hg18KG
+ cd /cluster/data/aplCal1/bed/tblastn.hg18KG
+ echo ../../blastDb/*.nsq | xargs ls -S | sed "s/\.nsq//" > query.lst
+ wc -l query.lst
+# 656 query.lst
+
+ # we want around 50000 jobs
+ calc `wc /cluster/data/hg18/bed/blat.hg18KG/hg18KG.psl | awk '{print $1}'`/\(50000/`wc query.lst | awk '{print $1}'`\)
+
+# 36727/(50000/656) = 481.858240
+
+ mkdir -p kgfa
+ split -l 482 /cluster/data/hg18/bed/blat.hg18KG/hg18KG.psl kgfa/kg
+ cd kgfa
+ for i in *; do
+ nice pslxToFa $i $i.fa;
+ rm $i;
+ done
+ cd ..
+ ls -1S kgfa/*.fa > kg.lst
+ wc kg.lst
+# 77 77 1001 kg.lst
+
+ mkdir -p blastOut
+ for i in `cat kg.lst`; do mkdir blastOut/`basename $i .fa`; done
+ tcsh
+ cd /cluster/data/aplCal1/bed/tblastn.hg18KG
+ 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=/hive/data/outside/blast229/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 /hive/data/outside/blast229/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" -nohead $f.3 /cluster/data/aplCal1/blastDb.lft carry $f.2
+ liftUp -nosort -type=".psl" -pslQ -nohead $3.tmp /cluster/data/hg18/bed/blat.hg18KG/protein.lft warn $f.3
+ if pslCheck -prot $3.tmp
+ then
+ mv $3.tmp $3
+ rm -f $f.1 $f.2 $f.3 $f.4
+ fi
+ exit 0
+ fi
+fi
+rm -f $f.1 $f.2 $3.tmp $f.8 $f.3 $f.4
+exit 1
+'_EOF_'
+ # << happy emacs
+ chmod +x blastSome
+ exit
+
+ ssh swarm
+ cd /cluster/data/aplCal1/bed/tblastn.hg18KG
+ gensub2 query.lst kg.lst blastGsub blastSpec
+ para create blastSpec
+# para try, check, push, check etc.
+
+ para time
+# Completed: 50512 of 50512 jobs
+# CPU time in finished jobs: 4656584s 77609.74m 1293.50h 53.90d 0.148 y
+# IO & Wait Time: 281459s 4690.98m 78.18h 3.26d 0.009 y
+# Average job time: 98s 1.63m 0.03h 0.00d
+# Longest finished job: 409s 6.82m 0.11h 0.00d
+# Submission to last job: 10255s 170.92m 2.85h 0.12d
+
+ ssh swarm
+ cd /cluster/data/aplCal1/bed/tblastn.hg18KG
+ mkdir chainRun
+ cd chainRun
+ tcsh
+ cat << '_EOF_' > chainGsub
+#LOOP
+chainOne $(path1)
+#ENDLOOP
+'_EOF_'
+
+ cat << '_EOF_' > chainOne
+(cd $1; cat q.*.psl | simpleChain -prot -outPsl -maxGap=12000 stdin ../c.`basename $1`.psl)
+'_EOF_'
+ chmod +x chainOne
+ ls -1dS ../blastOut/kg?? > chain.lst
+ gensub2 chain.lst single chainGsub chainSpec
+ # do the cluster run for chaining
+ para create chainSpec
+ para try, check, push, check etc.
+
+# Completed: 77 of 77 jobs
+# CPU time in finished jobs: 2940s 49.01m 0.82h 0.03d 0.000 y
+# IO & Wait Time: 5988s 99.79m 1.66h 0.07d 0.000 y
+# Average job time: 116s 1.93m 0.03h 0.00d
+# Longest finished job: 297s 4.95m 0.08h 0.00d
+# Submission to last job: 455s 7.58m 0.13h 0.01d
+
+ cd /cluster/data/aplCal1/bed/tblastn.hg18KG/blastOut
+ for i in kg??
+ do
+ cat c.$i.psl | awk "(\$13 - \$12)/\$11 > 0.6 {print}" > 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.*.psl m60* | uniq | sort -T /tmp -k 14,14 -k 16,16n -k 17,17n > ../blastHg18KG.psl
+ cd ..
+ pslCheck blastHg18KG.psl
+# checked: 5050 failed: 0 errors: 0
+
+ # load table
+ ssh hgwdev
+ cd /cluster/data/aplCal1/bed/tblastn.hg18KG
+ hgLoadPsl aplCal1 blastHg18KG.psl
+
+ # check coverage
+ featureBits aplCal1 blastHg18KG
+# 6922807 bases of 619228098 (1.118%) in intersection
+
+ featureBits aplCal1 blastHg18KG xenoRefGene -enrichment
+# blastHg18KG 1.118%, xenoRefGene 0.845%, both 0.490%, cover 43.82%, enrich 51.85x
+
+ rm -rf blastOut
+#end tblastn