# directories
PCEM-SCRIPTS = /projects/compbio/experiments/models.97/scripts
MOTIF-SCRIPTS = /projects/compbio/experiments/motif-models/scripts
PCEF-SCRIPTS = /projects/compbio/experiments/test-set/fold-recognition-test/scripts
PCEM-INDEXES = /projects/compbio/experiments/models.97/indexes
PCEM-PDB = /projects/compbio/experiments/models.97/pdb
LIB= /projects/compbio/lib

# programs in non-standard places
PHYTREE = /projects/compbio/usr/karplus/src/phytree/phytree
DG = /projects/compbio/usr/karplus/src/phytree/dg

# libraries and data files
NRP= /projects/compbio/data/nrp/NRP_ALL.PDB
PDB= /projects/compbio/data/pdb/all-protein
MIXTURE = ${LIB}/recode2.20comp
TRANS-REG = ${LIB}/fssp-trained.regularizer


TARGET= t49

# change the following two lines (or redefine when calling make)
# to use a different chain as a template.
PRED=3pte
PRED2=3p


# source directory and files for template
PRED-DIR = ${PCEM-PDB}/${PRED2}/${PRED}
PRED-SEQ = ${PRED-DIR}/info/${PRED}.hssp.guide
PRED-A2M = ${PRED-DIR}/nostruct-align/${PRED}.target98.a2m.gz 
PRED-FSSP-A2M = ${PRED-DIR}/struct-align/${PRED}.fssp-t98.a2m.gz 
PRED-CONST-A2M = ${PRED-DIR}/struct-align/${PRED}.constr-t98.a2m.gz 

# destination directory for alignments
PRED-DEST= ${PRED}

#destination file prefixes for joint alignments
PT = ${PRED-DEST}/${PRED}-${TARGET}
TP = ${PRED-DEST}/${TARGET}-${PRED}

default: ${TARGET}.all 
joints:
	-make -k PRED=3pte PRED2=3p joint
	-make -k PRED=2bltA PRED2=2b joint
	

joint: \
	${PT}-fssp-global.pw.a2m.gz \
	${PT}-const-global.pw.a2m.gz\
	${TP}.pw.a2m.gz ${PT}.pw.a2m.gz \
	${TP}-global.pw.a2m.gz 	${PT}-global.pw.a2m.gz 

#	${TP}--0.7.pw.a2m.gz ${PT}--0.7.pw.a2m.gz \
#	${TP}--0.85.pw.a2m.gz ${PT}--0.85.pw.a2m.gz \


.PRECIOUS: \
%.gz %.t98.a2m.gz %.t98_4.a2m.gz .t98_5.a2m.gz .t98_6.a2m.gz \
	%.dist %-varh50-pdb.rdb %-t98.rdb %-fssp-t98.rdb \
	%.tree %_sorted.a2m.gz %_sorted.ids %.tree.ps \
	%.nrp-ids 




%.all:
	make $*.t98_6.a2m.gz \
		$*.t98_4-varh50-pdb.rdb \
		$*.t98_5-varh50-pdb.rdb \
		$*.t98_6-varh50-pdb.rdb \
		$*-t98.rdb \
		$*-sum98.rdb \
		$*.t98_6.tree 

%.gz: %
	gzip -f $^

%.pa: %.a2m.gz
	prettyalign $^ -m5 > $@

# fimstrength for target98 (but not for joint models)
FIMSTR = 1.0

%.t98_4.a2m.gz %.t98_5.a2m.gz %.t98_6.a2m.gz: %.seq
	${PCEM-SCRIPTS}/target98-with-param \
		 -out $*.t98 \
		 -seed $^ \
		 -iter 6 -all \
		 -thresholds -50,-40,-30,-22,-16,-14 \
		 -build varh50 \
		 -weight_dist \
		 -fimstrength ${FIMSTR}
	gzip -f $*.t98_*.a2m

	
%-t98.rdb: %.seq
	${PCEM-SCRIPTS}/score-lib \
		-chains ${PCEM-INDEXES}/target98.ids \
		-lib pdb -lib domains -a2m target98 \
		-build varh50 -db $^ \
		-NLLminusNULL -4 \
		-NLLreverse 1.0 \
		-viterbi 0 -sw_score 2 \
		-output $@

%-fssp-t98.rdb: %.seq
	${PCEM-SCRIPTS}/score-lib \
		-chains ${PCEM-INDEXES}/fssp-t98.ids \
		-lib pdb -lib domains -a2m fssp-t98 \
		-build varh50 -db $^ \
		-NLLminusNULL -4 \
		-NLLreverse 1.0 \
		-viterbi 0 -sw_score 2 \
		-output $@

%-sum98.rdb: %-t98.rdb %.t98_6-varh50-pdb.rdb
	column -c HMM tt -c TARGET HMM SCORE \
		< $*.t98_6-varh50-pdb.rdb \
		| column -c tt TARGET HMM SCORE > tmp.rdb
	sed 's/T00/t/' < $*-t98.rdb > tmp2.rdb
	${PCEF-SCRIPTS}/combine-scores -sum tmp.rdb tmp2.rdb \
	| sorttbl SCORE \
	> $@
	-rm tmp.rdb tmp2.rdb

${PRED-DEST}:
	-mkdir ${PRED-DEST}
	-fixmode ${PRED-DEST}

${TP}.pw.a2m.gz: ${TARGET}.t98_6.a2m.gz \
	${PRED-SEQ} ${PRED-A2M}
	make ${PRED-DEST}
	${PCEM-SCRIPTS}/build-joint98 \
		-thresholds -9,-6,-2 \
		-seqA ${TARGET}.seq \
		-seqB ${PRED-SEQ} \
		-a2mA $< -a2mB ${PRED-A2M} \
		-mult ${TP}.mult \
		-dist ${TP}.dist \
		> ${TP}.pw.a2m
	gzip -f ${TP}.*

${PT}.pw.a2m.gz: ${TARGET}.t98_6.a2m.gz \
	${PRED-SEQ} ${PRED-A2M} 
	make ${PRED-DEST}
	${PCEM-SCRIPTS}/build-joint98 \
		-useB \
		-thresholds -9,-6,-2 \
		-seqA ${TARGET}.seq \
		-seqB ${PRED-SEQ} \
		-a2mA $< -a2mB ${PRED-A2M} \
		-mult ${PT}.mult \
		-dist ${PT}.dist \
		> ${PT}.pw.a2m
	gzip -f ${PT}.*
	
${TP}-%.pw.a2m.gz: ${TARGET}.t98_6.a2m.gz \
	${PRED-SEQ} ${PRED-A2M} 
	make ${PRED-DEST}
	${PCEM-SCRIPTS}/build-joint98 \
		-thresholds -9,-6,-2 \
		-seqA ${TARGET}.seq \
		-seqB ${PRED-SEQ} \
		-a2mA $< -a2mB ${PRED-A2M} \
		-mult ${TP}-$*.mult \
		-dist ${TP}-$*.dist \
		-fimstrength $* \
		> ${TP}-$*.pw.a2m
	gzip -f ${TP}-$*.*

${PT}-%.pw.a2m.gz: ${TARGET}.t98_6.a2m.gz \
	${PRED-SEQ} ${PRED-A2M} 
	make ${PRED-DEST}
	${PCEM-SCRIPTS}/build-joint98 \
		-useB \
		-thresholds -9,-6,-2 \
		-seqA ${TARGET}.seq \
		-seqB ${PRED-SEQ} \
		-a2mA $< -a2mB ${PRED-A2M} \
		-mult ${PT}-$*.mult \
		-dist ${PT}-$*.dist \
		-fimstrength $* \
		> ${PT}-$*.pw.a2m
	gzip -f ${PT}-$*.*

${TP}-global.pw.a2m.gz : ${TARGET}.t98_6-varh50.mod ${PRED-SEQ} 
	make ${PRED-DEST}
	hmmscore ${TP}-global.pw -i $< -db ${TARGET}.seq -db ${PRED-SEQ} \
		-viterbi 0 \
		-subtract_null 4 -simplethreshold 1 \
		-select_align 8
	gzip -f ${TP}-global.pw.a2m

${PT}-global.pw.a2m.gz : ${PRED-A2M:.a2m.gz=-varh50.mod} 
	make ${PRED-DEST}
	hmmscore ${PT}-global.pw -i $< -db ${TARGET}.seq -db ${PRED-SEQ} \
		-viterbi 0 \
		-subtract_null 4 -simplethreshold 1 \
		-select_align 8
	gzip -f ${PT}-global.pw.a2m

${PT}-fssp-global.pw.a2m.gz : ${PRED-FSSP-A2M:.a2m.gz=-varh50.mod} 
	make ${PRED-DEST}
	hmmscore ${PT}-fssp-global.pw -i $< -db ${TARGET}.seq -db ${PRED-SEQ} \
		-viterbi 0 \
		-subtract_null 4 -simplethreshold 1 \
		-select_align 8
	gzip -f ${PT}-fssp-global.pw.a2m

${PT}-const-global.pw.a2m.gz : ${PRED-CONST-A2M:.a2m.gz=-varh50.mod} 
	make ${PRED-DEST}
	hmmscore ${PT}-const-global.pw -i $< -db ${TARGET}.seq -db ${PRED-SEQ} \
		-viterbi 0 \
		-subtract_null 4 -simplethreshold 1 \
		-select_align 8
	gzip -f ${PT}-const-global.pw.a2m
	
%.dist: %.mod ${TRAIN_SET}
	hmmscore $* -i $*.mod -db ${TRAIN_SET} -sw 2 \
		-subtract_null 4 -simplethreshold 1 -sort 2 \
		-fimstrength ${FIMSTR}

%-varh50-pdb.rdb: %.a2m.gz
	${PCEM-SCRIPTS}/score-target -alignment $^ -build varh50 \
		-db ${PDB} \
		-NLLminusNULL	-5.0 -NLLreverse -1.0 \
		-viterbi 0 -sw_score 2 \
		-output $@

%-pdb.dist %-pdb.a2m: %.mod ${PDB}
	hmmscore $*-pdb -i $*.mod -db ${PDB} -sw 2 \
		-subtract_null 4 -simplethreshold 1 -sort 2 \
		-select_align 2 \
		-fimstrength ${FIMSTR}
	gzip -f $*-pdb.a2m


%-h0.5.mod: %.a2m.gz
	${PCEM-SCRIPTS}/h0.5 $^ $@

%-varh50.mod: %.a2m.gz
	${PCEM-SCRIPTS}/varh50 $^ $@
%.h1.0.mod: %.a2m.gz
	${PCEM-SCRIPTS}/h1.0 $^ $@
	
%_sorted.ids %.tree %_sorted.a2m.gz %.phytrace: %.a2m.gz
	gunzip $^
	${PHYTREE} -o -i -r flat $* $*.a2m ${LIB}/recode2.20comp
	gzip -f $*.a2m
	gzip -f $*_sorted.a2m
	-rm $*.phytrace $*.tree_weight

%.tree.ps: %.tree
	${DG} $^


%.ids: %.a2m.gz
	zcat $^ \
	| ids-from-fasta > $@

# make id list for sequences from nrp, stripping off multdomain range info
# (Assumes that a colon in a name indicates an NRP name).
%.nrp-ids: %.a2m.gz
	zcat $^ \
	| ids-from-fasta -nodom \
	| egrep ":" \
	> $@

%.seq.gz: %.ids
	extract-from-fasta $^ < ${NRP} > $*.seq
	gzip -f $*.seq



%.a2m-with-dots: %.a2m.gz
	zcat $^ > tmp
	prettyalign tmp -f > $@