diff --git a/projects/novelfams/translate2modelcif.py b/projects/novelfams/translate2modelcif.py
index 22b22b7fea948c3fff1ba70952abe27c88b8509c..8cfbf839eb16606be308e80ee7949e45b33df531 100644
--- a/projects/novelfams/translate2modelcif.py
+++ b/projects/novelfams/translate2modelcif.py
@@ -6,6 +6,7 @@
 
 from timeit import default_timer as timer
 import argparse
+import datetime
 import gzip
 import os
 import shutil
@@ -494,9 +495,45 @@ def _get_sequence_dbs_colabfold(seq_dbs):
     return [db_dict[seq_db] for seq_db in seq_dbs]
 
 
+def _get_sequence_dbs_alphafold(seq_dbs):
+    """Get AlphaFold seq. DBs."""
+    db_dict = {
+        "MGnify": modelcif.ReferenceDatabase(
+            "MGnify",
+            "https://storage.googleapis.com/alphafold-databases/"
+            + "casp14_versions/mgy_clusters_2018_12.fa.gz",
+            version="2018_12",
+            release_date=datetime.datetime(2018, 12, 6),
+        ),
+        "UniRef90": modelcif.ReferenceDatabase(
+            "UniRef90",
+            "ftp://ftp.uniprot.org/pub/databases/uniprot/uniref/uniref90/"
+            + "uniref90.fasta.gz",
+            version=None,
+            release_date=None,
+        ),
+        "BFD": modelcif.ReferenceDatabase(
+            "BFD",
+            "https://storage.googleapis.com/alphafold-databases/"
+            + "casp14_versions/"
+            + "bfd_metaclust_clu_complete_id30_c90_final_seq.sorted_opt.tar.gz",
+            version="6a634dc6eb105c2e9b4cba7bbae93412",
+        ),
+        "Uniclust30": modelcif.ReferenceDatabase(
+            "Uniclust30",
+            "https://storage.googleapis.com/alphafold-databases/"
+            + "casp14_versions/uniclust30_2018_08_hhsuite.tar.gz",
+            version="2018_08",
+            release_date=None,
+        ),
+    }
+    return [db_dict[seq_db] for seq_db in seq_dbs]
+
+
 def _get_modelcif_protocol_data(data_labels, target_entities, model, msa):
     """Assemble data for a ModelCIF protocol step."""
     data = modelcif.data.DataGroup()
+
     for label in data_labels:
         if label == "target_sequences":
             data.extend(target_entities)
@@ -506,6 +543,12 @@ def _get_modelcif_protocol_data(data_labels, target_entities, model, msa):
             data.extend(
                 _get_sequence_dbs_colabfold(["UniRef", "Environmental"])
             )
+        elif label == "alphafold_reference_dbs":
+            data.extend(
+                _get_sequence_dbs_alphafold(
+                    ["MGnify", "UniRef90", "BFD", "Uniclust30"]
+                )
+            )
         elif label == "msas":
             data.append(msa)
         else:
@@ -684,7 +727,6 @@ def _get_protocol_steps_and_software_colabfold(config_data):
     protocol = []
 
     # MSA step
-    # Step 1 - MSA: Using default Colabfold databases with default parameters (colabfold_envdb_202108, uniref30_2202)
     step = {
         "method_type": "coevolution MSA",
         "name": None,
@@ -738,30 +780,50 @@ def _get_config_colabfold():
 
 def _get_config_alphafold():
     """Get config variables for AlphaFold"""
-    description = "Predict model coordinates using AlphaFold."
+    af2_version = "2.2.0"
+    msa_description = (
+        "MSAs created for corresponding target sequence with AlphaFold using "
+        + "default parameters."
+    )
 
-    return {"description": description}
+    mdl_description = (
+        f"Model generated using AlphaFold ({af2_version} with default "
+        + "parameters) producing 5 models,ranked by pLDDT, starting from a the "
+        + f"Alphafold {af2_version} produced MSA."
+    )
+
+    return {
+        "af2_version": af2_version,
+        "msa_description": msa_description,
+        "mdl_description": mdl_description,
+    }
 
 
 def _get_protocol_steps_and_software_alphafold(config_data):
     """Get protocol steps for AF2 based models."""
     protocol = []
+    # MSA generation
+    step = {
+        "method_type": "coevolution MSA",
+        "name": None,
+        "details": config_data["msa_description"],
+        "input": ["target_sequences", "alphafold_reference_dbs"],
+        "output": ["msas"],
+        "software": [_get_af2_software(config_data["af2_version"])],
+        "software_parameters": None,
+    }
+    protocol.append(step)
 
     # modelling step
     step = {
         "method_type": "modeling",
         "name": None,
-        "details": config_data["description"],
+        "details": config_data["mdl_description"],
+        "input": ["target_sequences"],
+        "output": ["model"],
+        "software": [_get_af2_software("2.2.0")],
+        "software_parameters": None,
     }
-    # get input data
-    # Must refer to data already in the JSON, so we try keywords
-    step["input"] = ["target_sequences"]
-    # get output data
-    # Must refer to existing data, so we try keywords
-    step["output"] = ["model"]
-    # get software
-    step["software"] = [_get_af2_software("2.2.0")]
-    step["software_parameters"] = None
     protocol.append(step)
 
     return protocol