Question

about the runs with cn.mops

0

Entering edit mode

Bogdan ▴ 670

@bogdan-2367

Last seen 13 months ago

Palo Alto, CA, USA

Dear Gunter,

good morning and a happy autumn time ! when you have a minute, i would like to ask for an advice regarding the output of cn.mops on tumor-germline samples :

i've been using the algorithm that you have developed on approx 100 tumors that have a germline counterpart, and now I would like to better understand the results.

It would be great if you could tell me more about the differences between the functions :

-- segmentation()

-- cnvs()

-- cnvr()

For example, for chr21, after segmentation, we obtained the regions :

"seqnames" "start" "end" "width" "strand" "sampleName" "median" "mean" "CN"

"chr21" 1 12050000 12050000 "*" NA 0 0.0028 "CN2"

"chr21" 12050001 12140000 90000 "*" NA 1.0224545205372 2.1378 "CN3"

"chr21" 12140001 12870000 730000 "*" NA 0 -0.2329 "CN2"

"chr21" 12870001 46709983 33839983 "*" NA 1.19917489022848e-64 0.0135 "CN2

After cnvs() function :

"chr21" 12050001 12140000 90000 "*" "CN3

After cnvr() function :

"chr21" 12050001 12140000 90000 "*" "CN3"

And, what do "median" and "mean" fields represent, after segmentation() ?

It would be very helpful to hear from you, and many thanks in advance,

-- bogdan

cn.mops referencecn.mops • 1.7k views

ADD COMMENT • link 7.2 years ago Bogdan ▴ 670

0

Entering edit mode

Bogdan ▴ 670

@bogdan-2367

Last seen 13 months ago

Palo Alto, CA, USA

Dear Gunter,

please, it would be helpful, if I could have your insights on the following results. I have been running cnMOPS on TUMOR-GERMLINE sample, and for a particular case (on chr 11):

-- cnvs() function gives the results (below), that contain a combination of CN1, CN3, and CN4

chr11	124660001	126440000	1780000	*	NA	-1	-0.9855	CN1
chr11	126440001	127240000	800000	*	NA	0.5849625007	0.5846	CN3
chr11	127240001	128050000	810000	*	NA	0.5849625007	0.7208	CN3
chr11	128050001	130720000	2670000	*	NA	0.9999844562	0.8221	CN4
chr11	130720001	134030000	3310000	*	NA	0.5849625007	0.7167	CN3
chr11	134030001	134960000	930000	*	NA	1	1.0069	CN4
chr11	134960001	135086622	126622	*	NA	0.5849625007	0.6648	CN3

-- while for the same regions, cnvr() gives only 1 "concatenated" region, that is a CN1 :

chr11

124660001

135086622

10426622

*

CN1

ADD COMMENT • link 7.2 years ago Bogdan ▴ 670

1

Entering edit mode

Hello Bogdan,

You should go with the CNVs in the slot "cnvs()". The integer copy numbers in the slot "cnvr()" are the average integer copy numbers in the whole region. Since the region is larger than the individual CNVs, in fact the region is a union of all CNVs that occur there, the average copy number is typically CN2 again. What is reported in "cnvr()" is a kind of summary because users find it convenient to be pointed to locations in the genome, where copy number variations occur frequently. However, the actual results can be found in "cnvs()".

Regards,

Günter

ADD REPLY • link 7.2 years ago Günter Klambauer ▴ 540

0

Entering edit mode

Dear Gunter, thank you for your message and for confirmation. Yes, I thought that this may be the case, and we've used indeed the results from cnvs().

ADD REPLY • link 7.2 years ago Bogdan ▴ 670

score 3 · Accepted Answer · 2017-08-31

3

Entering edit mode

Günter Klambauer ▴ 540

@gunter-klambauer-5426

Last seen 3.9 years ago

Austria

Hello Bogdan,

Slot "segmentation": This contains the whole chromosome split into adjacent windows by breakpoints. This means that the initial windows, in which you counted the reads (e.g. 1000 bp), are joined to form longer segments. For each segment and sample, the inferred copy number is reported.
Slot "cnvs": A subset of segments from the ones from the segmentation slot. The segments that meet the criteria to be a copy number variation (CNV), i.e. being above or below the given thresholds, are contained in this slot.
Slot "cnvr": Whereas "segmentation" and "cnvs" report per sample, this slot reports CNVs per population. This means that this is a kind of single linkage clustering of the CNVs. Typically, CNVs of individuals occur frequently at similar locations in the genome. Therefore, I report the regions where any CNV occurs.

I hope this helps with the interpretation.

Regards,

Günter

ADD COMMENT • link 7.2 years ago Günter Klambauer ▴ 540

0

Entering edit mode

Dear Gunter, thanks a lot for a very prompt reply. If I may add another question please :

-- sometimes, in the file produced by the "segmentation" there are a few segments labelled as CN2 . For the case of tumor-normal comparison, what does it mean ? Is it related to potential LOH in the germline sample ?

-- and, how shall i interpret the "median" and "mean" numerical values, after segmentation() ? We are running the cn.mops algorithm on tumor-normal pairs, per chromosome.

thanks a lot,

-- bogdan

ADD REPLY • link 7.2 years ago Bogdan ▴ 670

1

Entering edit mode

The segmentation slot contains also normal ("CN2") segments, i.e. segment that do not contain a copy number variation. So, you should expect to find CN2 there.

The "median" and "mean" are the median and mean signed individual I/NI (sI/NI) calls across the joined segments (cn.mops supplies and sI/NI call for each of the initial segments). The higher the values, the higher the underlying copy number, the lower, the lower the copy number.

ADD REPLY • link 7.2 years ago Günter Klambauer ▴ 540

0

Entering edit mode

Dear Gunter, thank you again ;) I would like to better understand the differences between the results from CNVS() versus the results from CNVR(), in the context of CNV calling on TUMOR-GERMLINE, and if you do not mind providing some insights on another question below please.

ADD REPLY • link 7.2 years ago Bogdan ▴ 670