scripts/IsoSeq_ANGEL_IPS_commands.txt

Version 1

#InterProScan of Iso-seq transcripts not mapped (or ignored due to too many differences with genome) to the P citri genome
# precalc server lookup disabled because these are nonpublished sequences!
#run on bufo


mkdir /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS
mkdir /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS/input
mkdir /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS/output
mkdir /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS/scripts

cd /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS/scripts

conda activate isoseq

#Clone ANGEL repo and install

git clone https://github.com/Magdoll/ANGEL.git
cd ANGEL
# does not work with biopython version >1.77 !
conda install -n isoseq -c cctbx202008 biopython
python setup.py build
python setup.py install
cd ..

#run ANGEL - get best ORF from Iso-seq transcripts ## does not work due to problems with classifier_pickle file
#angel_predict.py \
#--output_mode=all \
#--min_angel_aa_length 100 \
#--min_dumb_aa_length 100 \
#--cpus 30 \
#../../_A_02_cDNAcupcake-dry/output/Pcitri.ignored_ids.fasta ./ANGEL/training_example/test_angel.ANGEL.pickle ../output/Pcitri.ignored_ids 


# outputs the longest ORF (or first ORF) that exceed the user-defined minimum length and have a positive log-odds scores based on hexamer frequencies
dumb_predict.py ../../_A_02_cDNAcupcake-dry/output/Pcitri.ignored_ids.fasta ../output/Pcitri.ignored_ids.dumb --min_aa_length 100 --cpus 30

conda deactivate

#switch to java 11
update-java-alternatives --list
sudo update-alternatives --set java /usr/lib/jvm/java-11-oracle
java -version

cd /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS/scripts

cat ../output/Pcitri.ignored_ids.dumb.final.pep | tr -d '*' > ../output/Pcitri.ignored_ids.dumb.final.pep.rmstar

#IPS run on translated aa (dumb predictions)
./interproscan-5.46-81.0/interproscan.sh \
--applications Pfam \
--seqtype p \
-i ../output/Pcitri.ignored_ids.dumb.final.pep.rmstar.IPS \
--output-dir ../output/ \
--cpu 30 \
--disable-precalc \
--formats TSV,GFF3,HTML

## training ANGEL (non-dumb) - NOT WORKING
mkdir /DATB/markop/_S_P4_Pcitri_IsoSeq/_A_03-IPS/output/trainingset
angel_make_training_set.py ../output/Pcitri.ignored_ids.dumb.final ../output/trainingset/Pcitri.ignored_ids.dumb.training --random --cpus 30
## failed
## running CD-HIT to generate non-redundant set....
## /bin/sh: cd-hit: command not found