Read alignment plot interpretation

[hoelzer]

Hi!

I wonder a bit on how to interpret ploidy based on the read alignment plots, for example:

Would this be a clear signal for diploid or triploid?

It's not the best Nanopore run and I restricted here to reads > 10kb so the coverage is not so high. Nevertheless, it seems okay w/ ~50X.

I also run nQuire which checks for ploidy based on SNPs and here the picture is quite clear towards triploid:

Also, I dont really get how, on the same contig, some positions can be e.g. diploid and others triploid?

Thanks!

[johnomics]

Hi Martin,

Thanks for using Tapestry. The ploidy estimates are quite rough, just a guide. By default, read depth is estimated in 10kb windows, and the median read depth for all windows across the genome is assumed to be the diploid read depth. Ploidy level is then calculated relative to that median/diploid read depth. So it's not very clever - it doesn't take into account SNP information for example. The nQuire result is more reliable.

Read depths are never very consistent as there are many things that can affect the number of reads sequenced from and aligning to a particular region of the genome. Your plot looks pretty normal to me, though the coverage dip in the middle (between 4-5 Mb) might be an interesting feature - a centromere, or a common repeat or misassembly perhaps.

But yes, the ploidy level isn't clear from the green lines at least. I would look at some of the other contigs. If most of your contigs have consistently lower read depths than this contig, I think that would be solid evidence for an increase in ploidy on this contig.

Best wishes John

[hoelzer]

Hi John,

Thanks a lot for the clarification! That helps regarding interpretation.

Thx for the tool and cheers, Martin

On Sat, 10 Jul 2021, 15:50 John Davey, @.***> wrote:

Hi Martin,

Thanks for using Tapestry. The ploidy estimates are quite rough, just a guide. By default, read depth is estimated in 10kb windows, and the median read depth for all windows across the genome is assumed to be the diploid read depth. Ploidy level is then calculated relative to that median/diploid read depth. So it's not very clever - it doesn't take into account SNP information for example. The nQuire result is more reliable.

Read depths are never very consistent as there are many things that can affect the number of reads sequenced from and aligning to a particular region of the genome. Your plot looks pretty normal to me, though the coverage dip in the middle (between 4-5 Mb) might be an interesting feature

• a centromere, or a common repeat or misassembly perhaps.

But yes, the ploidy level isn't clear from the green lines at least. I would look at some of the other contigs. If most of your contigs have consistently lower read depths than this contig, I think that would be solid evidence for an increase in ploidy on this contig.

Best wishes John

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/johnomics/tapestry/issues/7#issuecomment-877641204, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADN2CZZXHWKLB6LREM457J3TXBF3LANCNFSM5AARGQ5A .