When X.shage[1]>X.shape[0]

[DorothyWen1]

Hi there,

I am interested in using your program to analyse my data. However, when I used default setting, I get errors when the number of SNPs within the region is larger than the sample size. I checked regionsRanker.py. It looks to me that the problems are due to the below code. Could you please let me know how to fix this? Thanks.

def eigenDecompose(self, X, K, normalize=True):
if (X.shape[1] >= X.shape[0]):
s,U = la.eigh(K)

Cherry

[omerwe]

Hi Cherry, Thanks for trying out MKLMM! Can you please provide more details about the exact command you ran and the error message you received? Having a larger number of SNPs in a region than the sample size shouldn't lead to a problem (typically one region includes the entire genome), but I'm afraid it could use up a lot of memory, unless they are linear kernels. Is this related to what you're seeing? Thanks, Omer

…

On Wed, Sep 12, 2018 at 4:14 PM DorothyWen1 ***@***.***> wrote: Hi there, I am interested in using your program to analyse my data. However, when I used default setting, I get errors when the number of SNPs within the region is larger than the sample size. I checked regionsRanker.py. It looks to me that the problems are due to the below code. Could you please let me know how to fix this? Thanks. def eigenDecompose(self, X, K, normalize=True): if (X.shape[1] >= X.shape[0]): s,U = la.eigh(K) Cherry — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJog8xgeTxIiiG-2tKLX8e6TyPAt5hbBks5uaWssgaJpZM4WmFrT> .

[DorothyWen1]

Hi Omer,

Thanks for the prompt reply. I think I printed out the K before to try to figure out the errors. However, it says the K is empty. By simply looking at the code, it looks like the K is not defined (i.e. s,U = self.eigenDecompose(X[:, np.array(r)], None) in rankRegions). I might be wrong as I don't know python that well. The code I used is : python MKLMM rank_regions.py --bfile_train trainplink --pheno_train phenoplink --out RankRegion. Thanks so much!

Cherry.

[omerwe]

Hi Cherry, Can you successfully run the examples shown in the github page <https://github.com/omerwe/MKLMM>? If yes, then I would guess that there's some specific problem with your data, and it's very difficult for me to figure out what it is without more information... My best guess is that your input includes a region that doesn't have any SNPs (e.g., maybe all the SNPs were filtered out for some reason)? But I would need to see the data or to get more information (or at least to see the error output) to know more. Best, Omer

…

On Thu, Sep 13, 2018 at 12:24 AM DorothyWen1 ***@***.***> wrote: Hi Omer, Thanks for the prompt reply. I think I printed out the K before to try to figure out the errors. However, it says the K is empty. By simply looking at the code, it looks like the K is not defined (i.e. s,U = self.eigenDecompose(X[:, np.array(r)], None) in rankRegions). I might be wrong as I don't know python that well. The code I used is : python MKLMM rank_regions.py --bfile_train trainplink --pheno_train phenoplink --out RankRegion. Thanks so much! Cherry. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJog84dZgDqHfRZeAl1vyZINkMYwWo9iks5uad3fgaJpZM4WmFrT> .