NIPALS odd imputation values

[Former user]

Thanks again for your Melbourne workshop!

I've been running the NIPALS function on my own data. I have two datasets which are slightly different subsets of a larger dataset that have a large overlap and similar distributions (see summary 1 and 2 attached). When I run nipals on my second dataset I get reasonable imputed values however for the first dataset I get wildly out of range imputed values (see NIPALS 1 and 2 attached). There are quite a few more missing values in the first dataset, would this be what is causing the issue?

[Al J Abadi]

Hi, sorry we’re getting back to you now as we had not realised we have not updated our issues page on Bioconductor (we’re now on GitHub and discussions are on Discourse).

‌

it might very well be the case the case that you have too many missing values so you might wanna filter out a little before proceeding. We will be upgrading our available nipals methods so stay tuned! In the meantime, you might wanna check out a new method that has come out and might work better on your specific data. There is currently not an option to reconstruct your matrix though, as far as I know.