Translocation and Inversion

[andyjslee]

What are the best ways to encode a translocation and an inversion in .gfa file format? Could you please provide an example for each variant? Thank you!

[ekg]

For inversion, you need two paths. In the first region of the graph the orientation of the paths is the same. In the inversion region, one path switches orientation and traverses it in the reverse complement direction relative to the other. In the last region the orientations are the same. The inversion will form a bubble between the two flanking regions.

For translocation you would have regions of the graph with dissimilar or non homologous chromosomes represented as paths traversing the same region.

[andyjslee]

Thank you so much for your quick and detailed response. Based on your explanation, I've come up with small examples of each variant type. Would you say the following are valid examples of an inversion and a translocation?

In the inversion example, the chr1:11-20 segment is inverted (i.e. 5' CGACGACTAG 3' is the inversion variant) and both the reference segment and the inversion are represented.

In the translocation example, the translocation breakpoint (i.e. junction) is between chr1:55 and chr2:76.

#	inversion example
H	VN:Z:1.1
S	chr1-6-10 	AAAAA
S 	chr1-11-20	GATCAGCAGC
S 	chr1-21-25	TTTTT
L 	chr1-6-10 	+	chr1-11-20 	+
L 	chr1-11-20 	+ 	chr1-21-25 	+
P 	inv-1-1    chr1-6-10+,chr1-11-20+,chr1-21-25+
P 	inv-1-2    chr1-6-10+,chr1-11-20-,chr1-21-25+

#	translocation example
H	VN:Z:1.1
S 	chr1-51-55	CGCGT
S 	chr2-76-80	ACTGC
P 	tra    chr1-51-55+,chr2-76-80+

Additionally, if you were to implement the above in gfa v2.0, would you recommend using the group tag or the edge tag in place of the path tag?