A script for helping in clustering reference database for CDS annotation

Former-commit-id: 7babc60d47f433efd1301fbbe2a5714bfe7f7658
Former-commit-id: cf45c79769c6204598dd456573846496e4e834c0
This commit is contained in:
2022-02-14 15:10:47 +01:00
parent 05551549bf
commit 616fd2bb44

View File

@ -0,0 +1,73 @@
CDHIT_ID=0.7
CDHIT_DELTA=0.8
function clusterize() {
local prot="${1}"
local fastain="${prot}.fst"
local cdhitout="${prot}.cdhit"
local fasta1="${prot}.1l.fst"
rm -rf "${prot}"
mkdir -p "${prot}"
cd-hit -i "${fastain}" \
-o "${prot}/${cdhitout}" \
-c ${CDHIT_DELTA} \
-G 1 \
-g 1 \
-aL 0.95 \
-s ${CDHIT_ID} \
-b 350 -p 1 \
-d 0 -n 3
fasta1line "${fastain}" > "${prot}/${fasta1}"
pushd "$prot"
rm -rf "*.cluster.*.ids"
$AwkCmd -v prot="$prot" \
'/^>/ {cluster=$2} \
! /^>/ {sub("\\.\\.*$","",$3); \
filename=prot".cluster."cluster".ids"; \
print $3 >> filename ; \
close(filename) }' "${cdhitout}.clstr"
rm -f "../$prot.clean.fst"
for ids in *.cluster.*.ids ; do
cluster=$(printf "%03d" $(echo "${ids}" | $AwkCmd -F'.' '{print $3}'))
size=$(wc -l "$ids" | $AwkCmd '{print $1}')
fsize=$(printf '%05d' $size)
alignment="${prot}.cluster.${fsize}.${cluster}.fst"
if (( size > 1 )) ; then
egrep -f "$ids" -A 1 "${fasta1}" | \
egrep -v -- '^--$' | \
clustalo -i - > "$alignment"
else
egrep -f "$ids" -A 1 "${fasta1}" | \
egrep -v -- '^--$' > "$alignment"
fi
if (( size >= 5 )) ; then
egrep -f "$ids" -A 1 "${fasta1}" | \
egrep -v -- '^--$' | \
formatfasta >> "../$prot.clean.fst"
fi
rm -f "$ids"
done
rm -f "${fasta1}"
rm -f "${cdhitout}"
popd
}