40ed3e546ef868bffe4c338d93d6528138ecfc44
angie
  Wed Sep 30 09:25:12 2015 -0700
Add test cases (and bug fixes!) to make sure that we get the desired
behavior when an insertion is adjacent to a non-insertion, and when
an insertion falls at the start or end of the search region (if applicable):

- Include insertions that fall at the start or end of the search region

- If the primary row is an insertion, keep secondary non-insertion rows
to the left and right.

- If the primary row is a non-insertion, keep secondary insertion rows
at its start and end.

(i.e. keep insertions at boundaries -- but don't let any non-insertions
slip through)

The VCF logic is more complicated because VCF indels always include an
extra base to the left, so they appear to start before they actually do,
and can be interspersed with non-indels that start there.

diff --git src/hg/lib/bed.sql src/hg/lib/bed.sql
index 1b0ea8c..f8e1e14 100644
--- src/hg/lib/bed.sql
+++ src/hg/lib/bed.sql
@@ -1,14 +1,14 @@
 # bed.sql was originally generated by the autoSql program, which also 
 # generated bed.c and bed.h.  This creates the database representation of
 # an object which can be loaded and saved from RAM in a fairly 
 # automatic way.
 
 #Browser extensible data
 CREATE TABLE bed (
     chrom varchar(255) not null,	# Human chromosome or FPC contig
     chromStart int unsigned not null,	# Start position in chromosome
     chromEnd int unsigned not null,	# End position in chromosome
     name varchar(255) not null,	# Name of item
               #Indices
-    PRIMARY KEY(chrom,12)
+    PRIMARY KEY(chrom)
 );