Exact meaning of "Skipping very large window"

[Nils Kölling]

I am running Platypus 0.8.1 and getting quite a few of these messages:

Skipping very large window chr19:nn-nn of size 3682. Max window size is 2500 (set in option --maxSize)

In a different thread here you said that "If the region is skipped, no variants are called".

Could you please clarify whether this even applies to simple SNPs in this region? In other words, if Platypus is skipping a window from chr19:1-10000 and there is a SNP at chr19:5000, will this still get called?

My assumption was that skipping a window just means that a long variant is being ignored, but that any variant up to a size of --maxSize would always get called. However, reading your reply in the other thread makes me think that maybe no variants are getting called at all?

[Andy Rimmer]

Hi Nils,

If Platypus skips a window like this then no variants in the window are called. This normally only happens in regions which either have a very large variant candidate (which normally comes from the assembler) or which are very divergent from the reference. Are you using the assembler?

Kind regards,

Andy

--
You received this message because you are subscribed to the Google Groups "Platypus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to platypus-users+unsubscribe@googlegroups.com.
To post to this group, send email to platypus-users@googlegroups.com.
Visit this group at https://groups.google.com/group/platypus-users.
To view this discussion on the web, visit https://groups.google.com/d/msgid/platypus-users/bdff81f6-67cf-4a4a-aa80-1f7019b8eece%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Dr Andrew (Andy) Rimmer

[Nils K]

Hi Andy,

Thank you for your reply! Yes, I am using the assembler (--assemble=1).

So does this mean that using the assembler will result in variants being missed in cases where the assembly can result in too a large window?

What would you recommend as a solution for this? Call variants twice, once with and once without the assembler?

Best,
Nils

On Monday, 22 May 2017 11:15:01 UTC+1, Andy Rimmer wrote:

Hi Nils,

If Platypus skips a window like this then no variants in the window are called. This normally only happens in regions which either have a very large variant candidate (which normally comes from the assembler) or which are very divergent from the reference. Are you using the assembler?

Kind regards,

Andy

On Fri, May 19, 2017 at 5:22 PM, Nils Kölling <nkoe...@gmail.com> wrote:

I am running Platypus 0.8.1 and getting quite a few of these messages:

Skipping very large window chr19:nn-nn of size 3682. Max window size is 2500 (set in option --maxSize)

In a different thread here you said that "If the region is skipped, no variants are called".

Could you please clarify whether this even applies to simple SNPs in this region? In other words, if Platypus is skipping a window from chr19:1-10000 and there is a SNP at chr19:5000, will this still get called?

My assumption was that skipping a window just means that a long variant is being ignored, but that any variant up to a size of --maxSize would always get called. However, reading your reply in the other thread makes me think that maybe no variants are getting called at all?

--
You received this message because you are subscribed to the Google Groups "Platypus Users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to platypus-user...@googlegroups.com.
To post to this group, send email to platypu...@googlegroups.com.

Visit this group at https://groups.google.com/group/platypus-users.
To view this discussion on the web, visit https://groups.google.com/d/msgid/platypus-users/bdff81f6-67cf-4a4a-aa80-1f7019b8eece%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Andy Rimmer]

Hi Nils,

Yes it is quite possible that when the assembler finds a very large variant candidate, this can cause smaller variants to be missed, or can cause the windows to become too large. The integration between large and small variant calling is not as smooth as I would like.

I recommend a 2-pass approach if you are interested in large variants and small variants. The small variant calling will be perfectly good without the assembler, you might just get some false positives around the breakpoints of un-called large variants.

Kind regards,

Andy

To unsubscribe from this group and stop receiving emails from it, send an email to platypus-users+unsubscribe@googlegroups.com.
To post to this group, send email to platypus-users@googlegroups.com.

Visit this group at https://groups.google.com/group/platypus-users.

To view this discussion on the web, visit https://groups.google.com/d/msgid/platypus-users/9e99c771-a058-4855-a39d-c5c5b2d24a68%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Nils K]

Thank you Andy, that makes things clear! I will move to a 2-pass approach then.

Just as a feature suggestion:
I don't know about other users, but I would much prefer it if Platypus simply stopped if such an error occurred, rather than simply ignoring the region and carrying on. Otherwise it seems to me like it would be very easy to miss variants because of unexpected issues. Especially if you run Platypus in a pipeline and don't look at the log files manually every time.

Maybe there could be a --fail-on-error flag or something like that, which would cause Platypus to quit if it encounters such a problem? This could apply both to this issue, as well as when a window is skipped because of too many reads.

Best,
Nils

To view this discussion on the web, visit https://groups.google.com/d/msgid/platypus-users/9e99c771-a058-4855-a39d-c5c5b2d24a68%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.