fixed the feature where very long stretches of Ns aligned

src/umd/nucmer.cc

@@ -41,6 +41,27 @@ void reverse_complement(std::string& s) {
     *st = rc(*st);
 }
 
+
+void replace_n_random_letter(std::string& s) {
+  auto st = s.begin();
+  auto en = s.end() - 1;
+
+  for( ; st < en; ++st) {
+    if( *st == 'n' ){
+        int n=rand()%4; 
+        if(n==0){
+        *st='a';
+        }else if(n==1){
+        *st='c';
+        }else if(n==2){
+        *st='g';
+        }else{
+        *st='t';
+        }
+    }
+  }
+}
+
 void SequenceAligner::align(const char* query, size_t query_len, std::vector<postnuc::Alignment>& alignments) {
   std::vector<mgaps::Match_t>        fwd_matches(1), bwd_matches(1);
   FastaRecordSeq                     Query(query, query_len);
@@ -135,6 +156,7 @@ sequence_info::sequence_info(std::istream& data, size_t chunk_size)  {
     records.push_back({ sequence_offset, header_offset });
   }
   sequence += '`';
+  replace_n_random_letter(sequence);
   records.push_back({ sequence.size(), headers.size() });
 }