Hello, I am doing a DTU with DRIMseq and I can get the results according to the following model: design_full <- model.matrix(~condition, data=DRIMSeq::samples(d)) where "condition" is my variable of interest (factor with 4 levels) and executing in sequence d <- dmPrecision(d, design=design_full) d <- dmFit(d, design=design_full) d <- dmTest(d, coef="condition4") # condition4= one of the four levels against reference condition
However, now I would like to do the analysis for paired data (which would be more correct for this experimental design) and I have added the variable 'subject' to the model, which indicates the pairing of samples: design_full <- model.matrix(~condition+subject, data=DRIMSeq::samples(d)) In this case, however, the analysis stops and I get the following error messages:
Error: BiocParallel errors 1 remote errors, element index: 135 731 unevaluated and other errors first remote error: Error in optimHess(par = par, fn = dm_lik_regG, gr = dm_score_regG, x = x, : non-finite value supplied by optim
Do you have any suggestions for resolving these errors? Thank you in advance. Luca
Your code would be easier to understand if you add a small example of your dataset and if you correctly format the code
The dataset is made of 24 total samples from 6 different subjects analysed for 4 different conditions. Sample_ID= sample_1 to sample_24; subjects= subject_1 to subject_6; conditions= condition_1 to condition_4
The transcripts were quantified with Salmon and the object d was obtained as follows:
and then filtered as follows:
Then, I designed the model matrix:
and ran the analysis:
From here the error as reported above. Hope this is clear enough now. Thanks!
if it is helpful, I will add the dataset more explicitly:
Is there a solution/feedback on this issue? Thank you in advance!