Merge delimiter for input file to define how to number columns

[nroak]

I'm interested in defining column separator for -c option in bedtools merge. I have a bedj file where the 4th column of input file has characters that are automatically detected as column separators. I want to ignore those and use TAB as a column separator instead. I have given an example file below and the output I get vs output I expect.

#Input FIle (generated from bedtools intersect)
chr1	4510001	4769999	{"color":"rgba(128,0,128,1.00)","exon":[[4510001,4519999],[4760001,4769999]],"name":"_n19_qBL1.7390461E-4"}	.	-1	-1		.	0
chr1	4850001	5099999	{"color":"rgba(128,0,128,1.00)","exon":[[4850001,4874999],[5075001,5099999]],"name":"_n59_qBL1.1236174E-9"}	chr1	4857814	4897909	Tcea1	40095
chr1	4850001	5099999	{"color":"rgba(128,0,128,1.00)","exon":[[4850001,4874999],[5075001,5099999]],"name":"_n59_qBL1.1236174E-9"}	chr1	5070018	5162529	Atp6v1h	29981
chr1	4850001	5099999	{"color":"rgba(128,0,128,1.00)","exon":[[4850001,4874999],[5075001,5099999]],"name":"_n59_qBL1.1236174E-9"}	chr1	4909576	5070285	Rgs20	160709

# Expected output with bedtools merge -c 8,9 -o collapse,collapse
chr1	4510001	5099999 Tcea1,Atp6v1h,Rgs20 40095,29981,160709

# Actual Output
chr1	4850001	5099999	{"color":"rgba(128	0	128	5099999]],5099999]],5099999]]	"name":"_n59_qBL1.1236174E-9"}	chr1	4857814	4897909	Tcea1	40095,"name":"_n59_qBL1.1236174E-9"}	chr1	5070018	5162529	Atp6v1h	29981,"name":"_n59_qBL1.1236174E-9"}	chr1	4909576	5070285	Rgs20	160709

As you can see merge used comma as a delimiter instead of a TAB. A way to define this input delimiter would be extremely useful.