I have a gff file where I have already removed nested features, similar to the 1st, 2nd, 4th and 5th bed features as shown in the figure:

But now I want to remove features corresponding to 3, but in a way that I can keep the leftmost unique region of the first feature and rightmost unique region of the second feature, keeping the overlap as a part of the longer feature.

Example data:

scaffold1   RepeatMasker    similarity  1627986 1629296 11.5    +   .   Clust2783_Helitron
scaffold1   RepeatMasker    similarity  1628280 1638525 0   +   .   Clust1896_LTRRT
scaffold1   RepeatMasker    similarity  1634325 1644243 0   +   .   Clust1098_Helitron
scaffold1   RepeatMasker    similarity  1643445 1644561 2.3 +   .   Clust305_Helitron

Output:

scaffold1   RepeatMasker    similarity  1627986 1628279 11.5    +   .   Clust2783_Helitron
scaffold1   RepeatMasker    similarity  1628280 1638525 0   +   .   Clust1896_LTRRT
scaffold1   RepeatMasker    similarity  1638526 1644243 0   +   .   Clust1098_Helitron
scaffold1   RepeatMasker    similarity  1644244 1644561 2.3 +   .   Clust305_Helitron

Is there a simple way to do this. Please let me know if you have any further suggestions on how to remove such overlaps.

Thanks