interpretation of results / clustering with only 1 SV

[Sames-Jtudd]

Hi there, nice package! I just had a 2 quick questions. 1) in the cluster certainty file produced by the cluster step there are columns called average proportion(1/2). Could you elaborate of what the difference is between these two values and how they should best be resolved to a per SV CCF.

On the github page it states that this value (in the average proportion column) is the CCFpurity. Just wanted to clarify that this is correct. Unless mistaken the CCF (cancer cell fraction) purity = cellular fraction. So if you want the CCF you would need to convert it back (divide by the purity).

2) In samples with only one SV passing filtering the clustering step is not fully performed and doesn't produce a cluster certainty file. While clustering of 1 SVs in nonsensical it is relevant to cluster then with SNVs, ( i have provided an SNV file for this command) and also to produce an estimate of the CCF.

Is it possible to either force the program to perform clustering or generate a CCF estimates in these cases.

thanks

Jamie

[mcmero]

Hi Jamie,

1. average_proportion(1/2) are calculated using VAF(1/2) of the SV respectively (left/right normal read counts). See 'Calculating variant CCFs' section in the paper methods for more details. You are correct with the CCF calculation: average_proportion/purity will give you the CCF. Taking the average of CCF1 and CCF2 should be adequate for an SV-level CCF in most cases.

2. This is possible but currently not supported as ccube returns a slightly different format if you give it only one variant. I've made a note of this and will handle this case in a future release.

Cheers, Marek

[Sames-Jtudd]

Thanks for the quick reply.

Best Jamie

From: Marek Cmero notifications@github.com Sent: 26 March 2020 04:00 To: mcmero/SVclone SVclone@noreply.github.com Cc: James Studd James.Studd@icr.ac.uk; Author author@noreply.github.com Subject: Re: [mcmero/SVclone] interpretation of results / clustering with only 1 SV (#6)

Hi Jamie,

1. average_proportion(1/2) are calculated using VAF(1/2) of the SV respectively (left/right normal read counts). See 'Calculating variant CCFs' section in the paper methods for more details. You are correct with the CCF calculation: average_proportion/purity will give you the CCF. Taking the average of CCF1 and CCF2 should be adequate for an SV-level CCF in most cases.

2. This is possible but currently not supported as ccube returns a slightly different format if you give it only one variant. I've made a note of this and will handle this case in a future release.

Cheers, Marek

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/mcmero/SVclone/issues/6#issuecomment-604215256, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AICFCUNLYPDSFLUWNH5ADNDRJLHKHANCNFSM4LTQPLCQ.

The Institute of Cancer Research: Royal Cancer Hospital, a charitable Company Limited by Guarantee, Registered in England under Company No. 534147 with its Registered Office at 123 Old Brompton Road, London SW7 3RP.

This e-mail message is confidential and for use by the addressee only. If the message is received by anyone other than the addressee, please return the message to the sender by replying to it and then delete the message from your computer and network.

[mcmero]

The latest commit now handles the case of running a single SV through the clustering step, which can be used with the post-assign functionality.

[Sames-Jtudd]

Sweet, cheers that was fast.

Best

Jamie

From: Marek Cmero notifications@github.com Sent: 08 April 2020 02:42 To: mcmero/SVclone SVclone@noreply.github.com Cc: James Studd James.Studd@icr.ac.uk; Author author@noreply.github.com Subject: Re: [mcmero/SVclone] interpretation of results / clustering with only 1 SV (#6)

CAUTION: This email originated from outside of the ICR. Do not click links or open attachments unless you recognize the sender's email address and know the content is safe.

The latest commit now handles the case of running a single SV through the clustering step, which can be used with the post-assign functionality.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/mcmero/SVclone/issues/6#issuecomment-610704092, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AICFCUOMDQTMARPONQF7WKTRLPI6NANCNFSM4LTQPLCQ.

The Institute of Cancer Research: Royal Cancer Hospital, a charitable Company Limited by Guarantee, Registered in England under Company No. 534147 with its Registered Office at 123 Old Brompton Road, London SW7 3RP.

This e-mail message is confidential and for use by the addressee only. If the message is received by anyone other than the addressee, please return the message to the sender by replying to it and then delete the message from your computer and network.

[wanhui5867]

Hi Marek,

I also encountered only 1 SV after filtering. Is it different from other results in that it does not produce a graph pdf file??

Additionally, how about the situation that only 1 SV as raw SV input? I tested that it will arise Error TypeError: iteration over a 0-d array in the annotating step. I wonder if it is reasonable that let this situation also produce ccube results like the above situation?

[mcmero]

Thank you for raising this. I've just pushed a commit that deals with the situation of only 1 SV input at the annotate step.

Regarding the output plot, unfortunately this is not produced for a single SV, as this requires ccube's clustering results object, which is not available for a single SV.

[wanhui5867]

Thanks for your quick reply! I tried it and it works now.