JonathanShor / DoubletDetection

Doublet detection in single-cell RNA-seq data.
https://doubletdetection.readthedocs.io/en/stable/
MIT License
81 stars 23 forks source link

Reproducibility issue #135

Closed Jayjay601 closed 3 years ago

Jayjay601 commented 3 years ago

Hi,

I am using Reticulate in R to call for DoubletDetection. I set the random_state but the number of predicted doublets are different each time. So I did a few tries, either setting seed only in R, setting seed using py_set_seed() or both, with the random_state in DoubletDetection fixed.

seed <- as.integer(40) set.seed(seed) py_set_seed(seed, disable_hash_randomization = TRUE)

I set use_phenograph=FALSE. Is anyone else having the issue of reproducing the results? I don't know where the randomness is coming from

Thanks, J

EDIT: managed to reproduce the results.