We submitted both flanks for KG04117 and annotated it as a double
insert line.  However, the 5' flank sequence, which corresponds to
KG04417b (FBti0021399), was identical to that of KG04105 in an
adjacent well of the 96 well plate during the inverse PCR.  The 3'
flank of KG04105 mapped to the same site as the 5' flank.  In the
recheck of the final stock of KG04117, the sequence quality was low
for the 5' flank, but it gives a credible alignment to the same site
as the original (and recheck) 3' flank.  My reinterpretation is that
the original 5' flank of KG04117 was a contamination product from
KG04105.  Based on this interpretation, KG04117 should be annotated
as having a single insertion at the site in 43F that is now annotated
as KG04117a (FBti0021398).
By the same argument, KG05684a (FBti0024518) is also a cross-contamination
artifact and KG05684 should be annotated as a single insertion at the site
in Rab5 now annotated as KG05684b (FBti0024519).