Makefile as computational pipeline and multiple stems - makefile

Question

I'm writing a computational pipeline for some processing of neuroimaging data with make.
I chose to use make for this task for the advantages that it provides as incrementality, dependency management and easy parallelization of tasks (the latter two are my main focus).
I started writing the Makefile for a single subject and it worked correctly. Now I tried to extend it to perform the same tasks for all the subjects, but I got stuck on the commented part of the Makefile shown below.
File organization is like this:
aal/
data
subj1/
dti/
data
t1/
data
subj2/
dti/
data
t1/
data
subj3/
dti/
data
t1/
data
and this is what I currently got
LOGOPT := >> log.txt 2>> log.txt
SUBJS := subj1 subj2 subj3
MNI_ROIS := ${wildcard AAL/AAL_to_MNI_*.nii.gz}
DIFF_ROIS := ${subst AAL/AAL_to_MNI_,DTI/aal2diff/AAL_,$(MNI_ROIS)}
all: $(SUBJS)
$(SUBJS): % : %/transf/MNI_to_T1_warp.nii.gz %/transf/T1_to_DTI.mat ${addprefix %/,$(DIFF_ROIS)}
#DTI/aal2diff/AAL_%.nii.gz: AAL/AAL_to_MNI_%.nii.gz DTI/nodif_brain.nii.gz transf/MNI_to_T1_warp.nii.gz transf/T1_to_DTI.mat | transf DTI/aal2diff
#applywarp --in=$< --ref=$(word 2,$^) --warp=$(word 3,$^) --postmat=$(word 4,$^) --out=$# --interp=nn $(LOGOPT)
#fslmaths $# -thrP 50 -bin $# $(LOGOPT)
$(addsuffix /DTI/aal2diff, $(SUBJS)):
mkdir -p $#
$(addsuffix /transf/T1_to_DTI.mat, $(SUBJS)): %/transf/T1_to_DTI.mat : %/transf/DTI_to_T1.mat | %/transf
convert_xfm -inverse $< -omat $# $(LOGOPT)
$(addsuffix /transf/DTI_to_T1.mat, $(SUBJS)): %/transf/DTI_to_T1.mat : %/DTI/nodif_brain.nii.gz %/T1/T1_brain.nii.gz | %/transf
flirt -in $< -ref $(word 2,$^) -omat $# -searchrx -90 90 -searchry -90 90 -searchrz -90 90 -dof 6 -cost corratio $(LOGOPT)
$(addsuffix /transf/MNI_to_T1_warp.nii.gz, $(SUBJS)): %/transf/MNI_to_T1_warp.nii.gz : %/transf/T1_to_MNI_warp.nii.gz %/T1/T1.nii.gz | %/transf
invwarp --warp=$< --ref=$(word 2,$^) --out=$# $(LOGOPT)
$(addsuffix /transf/T1_to_MNI_warp.nii.gz, $(SUBJS)): %/transf/T1_to_MNI_warp.nii.gz : %/T1/T1.nii.gz %/transf/T1_to_MNI.mat | %/transf
fnirt --in=$< --aff=$(word 2,$^) --cout=$# --config=T1_2_MNI152_2mm $(LOGOPT)
$(addsuffix /transf/T1_to_MNI.mat, $(SUBJS)): %/transf/T1_to_MNI.mat : %/T1/T1_brain.nii.gz | %/transf
flirt -in $< -ref /usr/local/fsl/data/standard/MNI152_T1_2mm_brain -omat $# -searchrx -90 90 -searchry -90 90 -searchrz -90 90 -dof 12 -cost corratio $(LOGOPT)
$(addsuffix /transf, $(SUBJS)):
mkdir -p $#
$(addsuffix /T1/T1_brain.nii.gz, $(SUBJS)): %/T1/T1_brain.nii.gz : %/T1/T1.nii.gz
bet $< $# -B -f 0.3 -g 0 -m $(LOGOPT)
$(addsuffix /DTI/nodif_brain.nii.gz, $(SUBJS)): %/DTI/nodif_brain.nii.gz : %/DTI/data.nii.gz
bet $< $# -f 0.3 -g 0 -m $(LOGOPT)
$(addsuffix /DTI/data.nii.gz, $(SUBJS)): %/DTI/data.nii.gz : %/DTI/diffusion.nii.gz
eddy_correct $< $# 0 $(LOGOPT)
Intuitively, what I would like to have is the possibility of having multiple stems in the targets section, say, %1 and %2 in order to write something like
%1/DTI/aal2diff/AAL_%2.nii.gz: AAL/AAL_to_MNI_%2.nii.gz %1/DTI/nodif_brain.nii.gz %1/transf/MNI_to_T1_warp.nii.gz %1/transf/T1_to_DTI.mat | %1/transf %1/DTI/aal2diff
applywarp --in=$< --ref=$(word 2,$^) --warp=$(word 3,$^) --postmat=$(word 4,$^) --out=$# --interp=nn $(LOGOPT)
fslmaths $# -thrP 50 -bin $# $(LOGOPT)
AFAIK, multiple stems aren't possible.
Do you have any suggestion of how can I go on?
Is it too bad to use make with something other than building programs?