Mercurial > repos > bgruening > edta

annotate make_pan_library.sh @ 0:f1a157358d4d draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
author bgruening
date Sun, 16 Oct 2022 12:41:19 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
1 #!/usr/bin/env bash
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
2
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
3 known_te=($1)
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
4
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
5 echo 'get classification info and convert #unknown to #DNA/Helitron'
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
6 for j in *mod.EDTA.TElib.novel.fa; do
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
7 for i in `cat $j.real`; do
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
8 grep $i $j;
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
9 done| \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
10 perl -nle 's/#unknown/#DNA\/Helitron/; print $_' > $j.real.ori &
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
11 done
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
12
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
13 wait
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
14
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
15 echo 'aggregate novel TE libraries'
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
16 i=0
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
17 for j in *real.ori; do
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
18 i=$(($i+5000));
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
19 perl /EDTA/util/rename_TE.pl $j $i;
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
20 done > NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
21 perl /EDTA/util/rename_TE.pl NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw > NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw2
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
22 mv NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw2 NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
23
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
24 # wait
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
25
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
26 # echo 'remove redundant'
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
27 # nohup perl /EDTA/util/cleanup_nested.pl \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
28 # -in NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
29 # -cov 0.95 \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
30 # -minlen 80 \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
31 # -miniden 80\
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
32 # -blast /opt/conda/lib/python3.6/site-packages/Bio/Blast &
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
33
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
34 # ls
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
35 # wait
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
36
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
37 # echo 'remove a number of false TEs and rename IDs'
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
38 # RepeatMasker -pa 36 -q -no_is -norna -nolow -div 40 -lib rm.fa -cutoff 225 NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw.cln
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
39 # perl /EDTA/util/output_by_list.pl 1 NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw.cln 1 \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
40 # <(awk '{print $5}' NAM.EDTA1.8.0.EDTA.TElib.novel.fa.raw.cln.out|grep TE) -ex -FA | \
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
41 # perl /EDTA/util/rename_TE.pl - > NAM.EDTA1.8.0.EDTA.TElib.novel.fa
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
42
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
43 # echo 'make comprehensive TE library'
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
44 # cat $known_te NAM.EDTA1.8.0.EDTA.TElib.novel.fa > NAM.EDTA1.8.0.TE11122019.TElib.fa
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
45
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
46 # echo 'finished make_pan_library'
f1a157358d4d planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/edta commit 24cb0421330e54b144b3e6f1be4ae35ac0e48c1c
bgruening
parents:
diff changeset
47