Somatic CNA's with tumor purity

[yashcrux]

Hi

I had a question regarding how the tumor purity values are parsed while somatic CNA.
I understand its by using '-t'. I don't understand whether it takes it in percentage or decimal.

For example if the purity is 88%, it shuld be introduced as "-t 88" or "-t 0.88"

Thanks for the help!

[tobiasrausch]

The -t command-line parameter is indeed somewhat related to tumor purity, ploidy and subclonality. To have maximum flexibility, I so far opted for a simple command-line parameter that determines the min. copy-number shift for the segmentation and somatic classification. For instance, for a diploid tumor with 80% tumor purity, a clonal genomic rearrangement of copy-number 3 would be observed at CN 2.8 = 0.8 * 3 + 0.2 * 2 so you can put a very stringent threshold of -t 0.5. However, for a tetraploid tumor with 30% tumor purity, a clonal genomic rearrangement of copy-number 5 would be observed at CN 4.3 = 0.3 * 5 + 0.7 * 4 so you need to use a very lenient cutoff such as -t 0.2 to detect such a shift in copy-number from 4 to 5.

[yashcrux]

Hi Tobias.

Thanks for this explanation!
Also can you please help me understand the relation of CN with -t.

Like what range should I consider to set, in this case, -t 0.5 when CN 2.8 or the other case.
in my case:
I have 2 tumor subtypes :
1st type: the values of CN ranges from 2.5 to 2.95
2nd type: ...CN 3.39 to 3.96

[yashcrux]

HI Tobias

I read the explanation again and decided on the -t values for :
1st type - 0.5
2nd type - 0.3

I don't know if its correct or I am in the right direction of the understanding of your explanation

Thanks in advance