The data type for Kmeans

[egabal]

Hello .. I am trying to use Shambhala2 for normalizing a dataset and I made sure that the format is compatible with that of P0 and Q0 csv files. However, I have no idea why i keep getting that error. I checked the class of EXP on matlab and it shows as 'double'. Does it have smth to do with my data file or should I modify a command in the matlab script.

Would kindly guide me through that and thanks in advance

Esraa

[BorisovNM]

Dear Esraa, Could you please share with me your code and data, which produce such an error? Kind regards. Nikolay Borisov, Ph.D., Prof. Bioinformatics Lab Leader Vivan Therapeutics Ltd., London, UK Armenian Bioinformatics Institute, Yerevan

пт, 16 февр. 2024 г. в 02:39, egabal @.***>:

Screenshot.2024-02-15.at.5.35.07.PM.png (view on web) https://github.com/BorisovNM/Shambhala2/assets/143759952/c53f9904-a8eb-4a20-af82-21215a9ee8a3

Hello .. I ma trying to use Shambhala2 for normalizing dataset and I and sure that the format is compatible with that of P0 and Q0 csv files. However, I have no idea why i keep getting that error. I checked the class of EXP on matlab and it shows as 'double'. Does it have smth to do with my data file or should I modify a command in the matlab script.

Would kindly guide me through that and thanks in advance

Esraa

— Reply to this email directly, view it on GitHub https://github.com/BorisovNM/Shambhala2/issues/2, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARCHY5TPFIL3BZF7K4XIGXTYT2FALAVCNFSM6AAAAABDLCJR2KVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEZTONJWGQZTCMI . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[egabal]

Thanks so much! I shared a repository with you! However, it is worth mentioning that this is data of Mus musculus .. while I found that in the function literature, it was mentioned that the Q reference dataset is meant for human gene expression datasets and applicability to other organisms will need adjustments. .. awaiting your kind feedback and thanks for your help!

Best

Esraa

[BorisovNM]

The P and Q datasets are adapted for human gene repertoire. Did you use murine or human gene symbols in your profiles that need to be harmonized? Best Nikolay

пт, 16 февр. 2024 г. в 20:02, egabal @.***>:

Thanks so much! I shared a repository with you! However, it is worth mentioning that this is data of Mus musculus .. while I found that in the function literature, it was mentioned that the Q reference dataset is meant for human gene expression datasets and applicability to other organisms will need adjustments. .. awaiting your kind feedback and thanks for your help!

Best

Esraa

— Reply to this email directly, view it on GitHub https://github.com/BorisovNM/Shambhala2/issues/2#issuecomment-1948735547, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARCHY5QSIXRL3FUVSHLRIJLYT57H5AVCNFSM6AAAAABDLCJR2KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNBYG4ZTKNJUG4 . You are receiving this because you commented.Message ID: @.***>

[egabal]

I retrieved the gene names from Ensemble using the biomaRt so I do have the gene symbol of Murine but i randomly checked some genes between the datasets and found them .. at the beginning the original dataset of mine had all gene names in lowercase and I converted them to UPPER before running the analysis to be similar as that of the auxiliary and reference datasets

thanks and sorry for the inconvenience

Esraa

[BorisovNM]

Dear Esraa, Shambhala2 uses normalization in the log-space. So that the expression data should not have zero counts. To avoid it, please use the standard trick: just add 1 to every count for every gene and every sample. Nikolay

[egabal]

Alright .. Will give it a try .. thanks so much :)

Esraa

[egabal]

did not work out :( .. got the same error! sorry for the inconvenience

Esraa

[BorisovNM]

Hmm... Interesting... I used Shambhala2 extensively during the last three years but never encountered such an error. It was a spontaneous conversion of an EXP array into a complex form before (or when) calling the CuBlock function in MATLAB. The function real(EXP) solved the problem. Many thanks for pointing to a bug, which does not exist anymore. I updated the code at the git repo. Cheers, NIkolay

сб, 17 февр. 2024 г. в 03:48, egabal @.***>:

did not work out :( .. got the same error! sorry for the inconvenience

Esraa

— Reply to this email directly, view it on GitHub https://github.com/BorisovNM/Shambhala2/issues/2#issuecomment-1949495117, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARCHY5VVQS5CDTMGYCPJZXDYT7V5LAVCNFSM6AAAAABDLCJR2KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNBZGQ4TKMJRG4 . You are receiving this because you commented.Message ID: @.***>

[egabal]

Oh Great! Thanks so much for your kind patience in tracking the problem and for helping me! I will get the updated code and run it again

sorry for the inconvenience as well

Best Regards

Esraa

[egabal]

it worked!!!!!! thanks so much