question regarding input vcf

[lyj95618]

Hello,

I have a question regarding the input vcf. When the input vcf is unphased, can I just use phaser to perform the read-backed phasing and then take the haplotypic_count.txt for ASE analysis using phaser gene AE script?

From the 2016 paper, if I didn't misunderstand, it seems like I can use phaser for variant phasing. But in the tutorial "Generate ASE Data with phASER", the input vcf is a phased vcf. I am just wondering whether a phased vcf is a must.

Any help would be much appreciated!
Laur

[secastel]

Hi Laur,

You can absolutely use an unphased VCF as input. It will just reduce the maximum length (and number of variants) that can be summed across when generating the counts. The phaser gene AE script will automatically select the phased haplotype with the greatest read depth and use that to generate the counts.

• Stephane

[lyj95618]

Thanks for answering my question!

Just two more things:

1. when WGS/WES is unavailable, is it possible to use panel sequence VCF file as the input VCF? will it work if I am only interested in a particular gene?

2. Would you recommend to perform "remove duplicated read (e.g. Picard)" before inputting the bam to phaser?

Thanks,
Laur

[secastel]

If by panel you mean genotyping array, then yes it should work. Theoretically it will work with any genotyping platform, provided of course that the genotype calls are accurate.

As for duplicate reads, I'd suggest reading our 2015 paper, where we discussing duplicate reads (https://link.springer.com/article/10.1186/s13059-015-0762-6)

[lyj95618]

Thank you very much for the reply!