From d5717e82128f5604b821133749b919299d15248c Mon Sep 17 00:00:00 2001
From: Milot Mirdita <milot@mirdita.de>
Date: Wed, 10 Feb 2021 20:20:41 +0100
Subject: [PATCH] Add CDD to databases downloader #410

---
 data/workflow/databases.sh | 15 ++++++++++++++-
 src/workflow/Databases.cpp |  7 +++++++
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/data/workflow/databases.sh b/data/workflow/databases.sh
index 4665ec61a..cbe2bb7f2 100644
--- a/data/workflow/databases.sh
+++ b/data/workflow/databases.sh
@@ -193,6 +193,15 @@ case "${SELECTION}" in
         fi
         INPUT_TYPE="eggNOG"
     ;;
+    "CDD")
+        if notExists "${TMP_PATH}/msa.msa.gz"; then
+            downloadFile "https://ftp.ncbi.nih.gov/pub/mmdb/cdd/cdd.info" "${TMP_PATH}/version"
+            downloadFile "https://ftp.ncbi.nih.gov/pub/mmdb/cdd/fasta.tar.gz" "${TMP_PATH}/msa.tar.gz"
+        fi
+        INPUT_TYPE="FASTA_MSA"
+        FASTA_MSA_SED='s|\.FASTA||g'
+        FASTA_MSA_MSA2PROFILE_PAR="--skip-query"
+    ;;
     "Resfinder")
         if notExists "${TMP_PATH}/download.done"; then
             downloadFile "https://api.bitbucket.org/2.0/repositories/genomicepidemiology/resfinder_db/commit/master?fields=hash,date" "${TMP_PATH}/version"
@@ -302,9 +311,13 @@ case "${INPUT_TYPE}" in
         # shellcheck disable=SC2086
         "${MMSEQS}" tar2db "${TMP_PATH}/msa.tar.gz" "${TMP_PATH}/msa" --output-dbtype 11 ${THREADS_PAR}  \
             || fail "tar2db died"
+        if [ -n "${FASTA_MSA_SED}" ]; then
+            sed "${FASTA_MSA_SED}" "${TMP_PATH}/msa.lookup" > "${TMP_PATH}/msa.lookup_tmp"
+            mv -f "${TMP_PATH}/msa.lookup_tmp" "${TMP_PATH}/msa.lookup"
+        fi
         rm -f "${TMP_PATH}/msa.tar.gz"
         # shellcheck disable=SC2086
-        "${MMSEQS}" msa2profile "${TMP_PATH}/msa" "${OUTDB}" --match-mode 1 --match-ratio 0.5 ${THREADS_PAR} \
+        "${MMSEQS}" msa2profile "${TMP_PATH}/msa" "${OUTDB}" --match-mode 1 --match-ratio 0.5 ${FASTA_MSA_MSA2PROFILE_PAR} ${THREADS_PAR} \
             || fail "msa2profile died"
         if [ -n "${REMOVE_TMP}" ]; then
             # shellcheck disable=SC2086
diff --git a/src/workflow/Databases.cpp b/src/workflow/Databases.cpp
index beaeed786..d8fd16e35 100644
--- a/src/workflow/Databases.cpp
+++ b/src/workflow/Databases.cpp
@@ -123,6 +123,13 @@ std::vector<DatabaseDownload> downloads = {{
     "https://xfam.wordpress.com/2020/06/30/a-new-pfam-b-is-released",
     false, Parameters::DBTYPE_HMM_PROFILE, databases_sh, databases_sh_len,
     { }
+}, {
+    "CDD",
+    "Conserved Domain Database is a protein annotation resource consisting of well-annotated MSAs for ancient domains and full-length proteins.",
+    "Lu et al: CDD/SPARCLE: the conserved domain database in 2020. Nucleic Acids Res 48(D1), D265–D268 (2020)",
+    "https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml",
+    false, Parameters::DBTYPE_HMM_PROFILE, databases_sh, databases_sh_len,
+    { }
 }, {
     "eggNOG",
     "eggNOG is a hierarchical, functionally and phylogenetically annotated orthology resource",