Error: Incompatibility between dimensions when running big_prodMat()

[nscarene]

Good morning Dr. Privé,

I hope this finds you well. I have an issue and was hoping you could help.
When I try to run the following line: pred_grid <- big_prodMat(G, beta_grid, ind.row = ind.val2, ind.col = df_beta[["NUM_ID"]]);
I automatically get an error message saying "error: you can't have missing values in 'X').

I tried using the snp_fastImputeSimple(G) to impute missing values but then I get another error message "error: Incompatibility between dimensions".

Could you please provide some insight on how to manage the error messages?

Thank you very much!

With kind regards,

[privefl]

How to handle missing values is a pinned issue here.

Just running snp_fastImputeSimple(G) gives you the error of incompatible dimensions? This seems hard to believe. Please provide a bit more details.

[nscarene]

Good morning Dr.

Thank you for your prompt response and my apologies for starting an issue that was already mentioned before.

I ran snp_fastImputeSimple(G) and then reran pred_grid <- big_prodMat(G, beta_grid, ind.row = ind.val2, ind.col = df_beta[["NUM_ID"]])

The error went from "Cannot have missing values in 'X'" before the imputation to "imcompatibility between dimensions". Should I rerun the analysis from the beginning, after that?

Also I was hoping you could help with a concern I have. I tried calculation LD correlation genome-wide using corr0 <- snp_cor(G, infos.pos = 1000 * CHR + POS2, size = 3 / 1000, ncores = NCORES). Is it normal that the operation takes several hours? Or is something wrong with the way I do it?

Any insight will be greatly appreciated.

With regards,

[privefl]

Please try to keep issues separated; open a new one if you need to.

Let's focus on the error of incompatible dimensions here.
Is it when running G2 <- snp_fastImputeSimple(G)? Or after running big_prodMat(G2, ...)?

[nscarene]

My apologies, I just submitted a new issue for the second question.

The error of incompatible dimensions arose after running big_prodMat(G2, ...) with G2 <- snp_fastImputeSimple(G)

[privefl]

It probably means that nrow(beta_grid) is not the same as length(df_beta[["NUM_ID"]]).

[nscarene]

I just checked and they are the same. See below,

nrow(beta_grid)
[1] 108372

length(df_beta[['NUM_ID']])
[1] 108372

[privefl]

Please reinstall remotes::install_github('privefl/bigassertr').
You should now get a more explicit error message if you rerun the function.

[nscarene]

Thank you do much Dr.

[nscarene]

UPDATE

using G2 <- snp_fastImputeSimple(G) helped take care of the error message.

Thank you!