How can we use a read simulator to generate paired-end data with uneven coverage?

Question

[zhangz.cs]

We plan to test a sv detection tool on paired-end data with uneven coverage. Could anyone please tell me how the data is generated?

Cheers,
Zhen

[GouthamAtla]

are you looking for only artificial data or also SNPs ?

[zhangz.cs]

I'm looking for only artificial data.

Answer 1

[thackl]

I've got an idea, although it might not be the most elegant way...

Simulators like ART can also produce SAM files for the simulated reads. If you convert these files to sorted BAMs, you could sample different fractions of reads from different regions of the reference and merge them into a unevenly covered fastq file.