Add AFDB50 to databases

steineggerlab · Aug 1, 2022 · 915ef7d · 915ef7d
1 parent 8b3b56e
commit 915ef7d
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 0 deletions.
diff --git a/data/structdatabases.sh b/data/structdatabases.sh
@@ -86,6 +86,15 @@ case "${SELECTION}" in
         push_back "${TMP_PATH}/afdb"
         INPUT_TYPE="FOLDSEEK_DB"
     ;;
+    "Alphafold/UniProt50")
+        if notExists "${TMP_PATH}/afdb50.tar.gz"; then
+            downloadFile "https://foldseek.steineggerlab.workers.dev/afdb50.tar.gz" "${TMP_PATH}/afdb50.tar.gz"
+            downloadFile "https://foldseek.steineggerlab.workers.dev/afdb50.version" "${TMP_PATH}/version"
+        fi
+        tar xvfz "${TMP_PATH}/afdb50.tar.gz" -C "${TMP_PATH}"
+        push_back "${TMP_PATH}/afdb50"
+        INPUT_TYPE="FOLDSEEK_DB"
+    ;;
     "Alphafold/Proteome")
         if notExists "${TMP_PATH}/alphafolddb.tar.gz"; then
             downloadFile "https://foldseek.steineggerlab.workers.dev/alphafolddb.tar.gz" "${TMP_PATH}/alphafolddb.tar.gz"

diff --git a/src/foldseek.cpp b/src/foldseek.cpp
@@ -227,6 +227,14 @@ std::vector<DatabaseDownload> externalDownloads = {
                 true, Parameters::DBTYPE_AMINO_ACIDS, structdatabases_sh, structdatabases_sh_len,
                 {}
         },
+        {
+                "Alphafold/UniProt50",
+                "AlphaFold UniProt Protein Structure Database clustered at 50% sequence identity.",
+                "Jumper et al. Highly accurate protein structure prediction with AlphaFold. Nature, (2021)",
+                "https://alphafold.ebi.ac.uk/",
+                true, Parameters::DBTYPE_AMINO_ACIDS, structdatabases_sh, structdatabases_sh_len,
+                {}
+        },
         {
                 "Alphafold/Proteome",
                 "AlphaFold Proteomes Protein Structure Database.",