Avoid trying to re-align and normalize very long indels which cause c…

…ause null pointer errors. These are dealt with by structural variation comparions. Thanks to Severine Catreux

project.clj

@@ -1,4 +1,4 @@
-(defproject bcbio.variation "0.1.6"
+(defproject bcbio.variation "0.1.7-SNAPSHOT"
   :description "Toolkit to analyze genomic variation data, built on the GATK with Clojure"
   :license {:name "MIT" :url "http://www.opensource.org/licenses/mit-license.html"}
   :dependencies [[org.clojure/clojure "1.5.1"]
@@ -21,8 +21,8 @@
                  [org.simpleframework/simple-xml "2.0.4"]
                  [org.apache.servicemix.bundles/org.apache.servicemix.bundles.jets3t "0.8.1_1"]
                  ;;
-                 [org.biojava/biojava3-core "3.0.4"]
-                 [org.biojava/biojava3-alignment "3.0.4"]
+                 [org.biojava/biojava3-core "3.0.8"]
+                 [org.biojava/biojava3-alignment "3.0.8"]
                  [org.clojars.chapmanb/circdesigna "0.0.2" :exclusions [net.sf.beaver/beaver-ant]]
                  [nz.ac.waikato.cms.weka/weka-stable "3.6.6"]
                  [org.clojars.chapmanb/fast-random-forest "0.98"]

src/bcbio/variation/complex.clj

@@ -30,8 +30,9 @@
 (defn- get-vc-alleles [vc]
   (vec (map #(.getDisplayString %) (cons (:ref-allele vc) (:alt-alleles vc)))))
 
-(defn is-multi-indel?
-  "Identify complex indels that can be split into multiple calls."
+(defn- is-multi-indel?
+  "Identify complex indels that can be split into multiple calls.
+   Caps indels we can operate on at 5000bp to avoid realignment errors for longer."
   [vc]
   (letfn [(monomorphic-alleles? [vc]
             (= 1 (->> (get-vc-alleles vc)
@@ -44,8 +45,12 @@
                  (not (monomorphic-alleles? vc))))
           (has-ref-padding-mismatch? [vc]
             (let [alleles (get-vc-alleles vc)]
-              (not= (nth (first alleles) 0) (nth (second alleles) 0))))]
+              (not= (nth (first alleles) 0) (nth (second alleles) 0))))
+          (splittable-size? [vc]
+            (< (apply max (map count (get-vc-alleles vc)))
+               5000))]
     (and (= "INDEL" (:type vc))
+         (splittable-size? vc)
          (or (has-multiple-nonref-alleles? vc)
              (has-ref-padding-mismatch? vc)))))