Using CPTAC dataset

[aarmey]

Using CPTAC dataset

[aarmey]

https://cptac-data-portal.georgetown.edu/study-summary/S046

[aarmey]

@mcreixell have you looked into this? If not I was thinking of doing some exploring about how to start this analysis...

[mcreixell]

Yes, I'm currently working on this.

[aarmey]

Ok great.

[aarmey]

It looks to me like the data is under the phosphoproteome folders here:

https://cptc-xfer.uis.georgetown.edu/publicData/Phase_III_Data/CPTAC_LUAD_S046/

You then go into the *PSM folder, then *.tsv folder, then all the files inside have the relevant data. Seems like a ton.

[aarmey]

Hmm... actually, not sure I see the location assignment within the protein. Must be around somewhere, though.

[mcreixell]

Yes, the number at the beginning of each file name (1-25) indicates the TMT10-plex experiment. Within each experiment, 8 are tumor-normal pairs, one is an internal reference (tumor or normal only), and the 10th is a pooled reference sample. I think all these files are integrated in the file I sent in my email via google drive.

[aarmey]

Aha, I see now.

[mcreixell]

The model seems to work. It's been running for ~2h and still active. How would you deal with the fact that the dataset contains tumor and healthy peptide signal for each patient? Would you model the entire data set as it is or maybe normalize to healthy?

[aarmey]

Wonderful! Might try profiling as the slow spots could be different. I'd suggest ignoring the healthy/normal labels for now and looking at them after clustering.