Add CATH50 #232

steineggerlab · Jan 26, 2024 · 6893dcc · 6893dcc
1 parent bb09017
commit 6893dcc
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 1 deletion.
diff --git a/data/structdatabases.sh b/data/structdatabases.sh
@@ -150,9 +150,17 @@ case "${SELECTION}" in
         push_back "${TMP_PATH}/pdb"
         INPUT_TYPE="FOLDSEEK_DB"
     ;;
+    "CATH50")
+        if notExists "${TMP_PATH}/cath50.tar.gz"; then
+            downloadFile "https://foldseek.steineggerlab.workers.dev/cath50.tar.gz" "${TMP_PATH}/cath50.tar.gz"
+            downloadFile "https://foldseek.steineggerlab.workers.dev/cath50.version" "${TMP_PATH}/version"
+        fi
+        tar xvfz "${TMP_PATH}/cath50.tar.gz" -C "${TMP_PATH}"
+        push_back "${TMP_PATH}/cath50"
+        INPUT_TYPE="FOLDSEEK_DB"
+    ;;
 esac
 
-
 if notExists "${OUTDB}.dbtype"; then
 case "${INPUT_TYPE}" in
     "FOLDSEEK_DB")

diff --git a/src/FoldseekBase.cpp b/src/FoldseekBase.cpp
@@ -397,5 +397,13 @@ std::vector<DatabaseDownload> externalDownloads = {
                 "https://www.rcsb.org",
                 true, Parameters::DBTYPE_AMINO_ACIDS, structdatabases_sh, structdatabases_sh_len,
                 {}
+        },
+        {
+                "CATH50",
+                "CATH domain database (combined AlphaFold and PDB CATH clustered at 50% seq.id.).",
+                "Bordin et al. AlphaFold2 reveals commonalities and novelties in protein structure space for 21 model organisms. Communications Biology, 6, 160 (2023)",
+                "https://www.cath.info",
+                true, Parameters::DBTYPE_AMINO_ACIDS, structdatabases_sh, structdatabases_sh_len,
+                {}
         }
 };