search

egr95 / R-codacore

An R package for learning log-ratio biomarkers from high-throughput sequencing data.
Other
21 stars 3 forks source link

type argument not working #5

Closed ggloor closed 1 year ago

ggloor commented 3 years ago

Hi

I'm trying this tool out on a couple of datasets and this is happening:

model.24=codacore(t(d.24), tf.conds, type = "balances" ) Error in codacore(t(d.24), tf.conds, type = "balances") : unused argument (type = "balances")

If I leave the type argument blank it works just fine

egr95 commented 3 years ago

Hi Greg,

Thanks a lot for pointing this out - the argument should read logRatioType = "balances", instead of just type = "balances". I had updated this in the code and in the guide, but not in the README, which I presume is where you saw this signature. Fixed now.

Thanks!

P.S: if you're interested in learning predictive Summed-Log-Ratios as well as balances, try logRatioType = "SLR". Also, I notice you are passing a response vector named tf.conds. Might this name be because the response vector is a Tensorflow object? This is totally fine, but note that this vector can just be a regular numeric vector (binary or continuous). In other words, it should not need to be converted to Tensorflow in order to make things work. I'm sure you're already aware, but just in case!