Cyril Matthey-Doret

Thank you for all the advice ! I tried running again with L=33 and U=5000. Indeed this increases the estimated ploidy level to 10, with 107 predicted chromosomes:  By...

As you suggested, I tried removing the hardcoded threshold (in PR #68) but it doesn't seem to be much better. Here is an example with `smudgeplot_plot.R -i "kmcdb_L50_U5000_coverages.tsv" -o "smudgeplot"...

Hi @davek44 thanks you are right, I can run the tutorial normally with the demo data. I think this is because I provided my own cool file and genome, so...

Thanks for your help, @davek44, My goal was to train the model on yeast. Because the genome is much smaller, I would need to reduce `-l` and probably the strides...

Indeed, I was referring to `akita_data.py`. That's correct, I can generate TFRecords, but `akita_train.py` crashes with the aforementioned `TypeError`. I did set `-l` to 131,072 instead of 1,048,576, as several...

Upgrading Renku did not solve the issue. I tried on another Ubuntu machine: same docker/renku/python versions and could not reproduce the issue. The only difference I could find is that...

ubuntu 20.04, docker 20.10.12

I see, thanks for the clarification. Whenever a dataset is imported from a specific (external) version, renku creates a _tag_ corresponding to that _version_, is that correct ? If so,...

For issue (2), I got it working with rpy2 version 2.9.0. I think this attribute was deprecated in 3.0 :)

The matrix is a bit sparse indeed, it has 258M contacts and below are pictures of the region chr3:160Mb-195Mb at 10 and 100 kb resolution respectively. The 100kb matrix looks...