diff --git a/schema/schema.sql b/schema/schema.sql
index 20c5f42..302d8cf 100755
--- a/schema/schema.sql
+++ b/schema/schema.sql
@@ -1,15 +1,19 @@
 
 CREATE EXTENSION IF NOT EXISTS pgcrypto;
 
+-- Data graph nodes
+
 CREATE TABLE IF NOT EXISTS backend_data_nodes (
-    item_id UUID NOT NULL DEFAULT gen_random_uuid() PRIMARY KEY,
-    metadata jsonb,
-    content VARCHAR,
-    type SMALLINT,
+    item_id    UUID DEFAULT gen_random_uuid() PRIMARY KEY,
+    metadata   jsonb NOT NULL,
+    content    VARCHAR NOT NULL,
+    type       SMALLINT NOT NULL,
     created_at TIMESTAMP DEFAULT NOW(),
-    seen BOOLEAN DEFAULT FALSE
+    has_phase  BOOLEAN DEFAULT FALSE -- TODO might want to reference the particular phase instead
 );
 
+-- Data graph edges
+
 CREATE TABLE IF NOT EXISTS backend_data_links (
     source_id UUID NOT NULL,
     target_id UUID NOT NULL,
@@ -18,14 +22,29 @@ CREATE TABLE IF NOT EXISTS backend_data_links (
     FOREIGN KEY (target_id) REFERENCES backend_data_nodes (item_id)
 );
 
-CREATE TABLE IF NOT EXISTS distinct_phases (
-    phid         INT PRIMARY KEY,
-    elements     VARCHAR(64) NOT NULL,
-    formula_txt  VARCHAR(128) NOT NULL,
-    formula_html VARCHAR(384) NOT NULL,
-    spg          INT,
-    pearson      VARCHAR(8),
-    crsystem     SMALLINT NOT NULL
+-- Pre-defined materials phases (read-only)
+
+CREATE TABLE IF NOT EXISTS backend_phases (
+    phase_id INT PRIMARY KEY,
+    elements VARCHAR(64) NOT NULL, -- FIXME?
+    formula  VARCHAR(128) NOT NULL,
+    spg      SMALLINT NOT NULL,
+    natcell  SMALLINT NOT NULL
+);
+
+CREATE INDEX IF NOT EXISTS i_phid ON backend_phases USING btree( phase_id );
+CREATE INDEX IF NOT EXISTS i_lattices ON backend_phases USING btree( spg );
+CREATE INDEX IF NOT EXISTS i_elements ON backend_phases USING btree( elements text_pattern_ops );
+
+-- Phase matching dIs
+
+CREATE TABLE IF NOT EXISTS backend_refdis (
+    ext_id   VARCHAR(128) NOT NULL,
+    phase_id INT NOT NULL,
+    provider SMALLINT DEFAULT 0, -- 0 Metis; 1 COD; 2 ICDD PDF4+; 3 CCDC; 4 MPDS;
+    elements VARCHAR(64) NOT NULL, -- FIXME?
+    di       NUMERIC[][] NOT NULL,
+    FOREIGN KEY (phase_id) REFERENCES backend_phases (phase_id)
 );
-CREATE INDEX IF NOT EXISTS i_phid ON distinct_phases USING btree( phid );
-CREATE INDEX IF NOT EXISTS i_elements ON distinct_phases USING btree( elements text_pattern_ops );
+
+CREATE INDEX IF NOT EXISTS i_elements ON backend_refdis USING btree( elements text_pattern_ops );