Separate complex nucleotide variants from other variants

chrovis · May 20, 2019 · 0270bf5 · 0270bf5
1 parent 7b324be
commit 0270bf5
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 7 deletions.
diff --git a/src/cljam/io/vcf/util.clj b/src/cljam/io/vcf/util.clj
@@ -328,7 +328,7 @@
        :ref (subs ref left-match (- ref-length right-match)),
        :alt (subs alt left-match (- alt-length right-match))}
 
-      :else {:type :other})))
+      :else {:type :complex})))
 
 (defn inspect-allele
   "Inspects an `alt` allele by comparing to a `ref` allele string.
@@ -342,7 +342,8 @@
   - `:deletion` Deletion of a short base sequence
   - `:complete-insertion` Complete insertion of a long sequence
   - `:breakend` Breakend of a complex rearrangement
-  - `:other` The variant is too complex or malformed"
+  - `:complex` Complex nucleotide variants other than snv/mnv/indel
+  - `:other` Can't categorize the allele, might be malformed"
   [ref alt]
   (let [alt-length (count alt)]
     (cond

diff --git a/test/cljam/io/vcf/util_test.clj b/test/cljam/io/vcf/util_test.clj
@@ -400,9 +400,9 @@
 
     "T" "<ctg1>CT" {:type :other} ;; invalid
     "T" "<CT" {:type :other} ;; invalid
-
     "." "A" {:type :other}
-    "TAC" "GC" {:type :other}
-    "TG" "TAC" {:type :other}
-    "TCA" "GGGG" {:type :other}
-    "AAAA" "CAC" {:type :other}))
+
+    "TAC" "GC" {:type :complex}
+    "TG" "TAC" {:type :complex}
+    "TCA" "GGGG" {:type :complex}
+    "AAAA" "CAC" {:type :complex}))