Attendants: 9
- Fixed a problem with running rstudio-server via singularity's rocker/rstudio image. Misses libgpng and libgeos. Dave Tang has an image here that contains libpng, but misses libgeos: https://davetang.org/muse/2021/04/24/running-rstudio-server-with-docker/ Philipp made a version that includes libgeos so Seurat can be installed: https://hub.docker.com/repository/docker/philippbayer/rstudio/ (should be philippbayer/rstudio:latest which will install v4.0.5)
- Fixed a barchart to add errorbars and use nicer scaling via scales:label_number()
- discussed Bayesian optimisation for scikit-learn
Attendants: 8
-
Fixed problems with
{glmnet}
packackage by converting input matrices from character matrices to numeric matrices- suggested to move the analysis to tidymodels at some point - support exists https://parsnip.tidymodels.org/reference/glmnet-details.html
-
learnt how to test for vectors being in other vectors in julia using
in()
in.(data.column, [vectors])
to create logical vector to subsetdata
rows- important to broadcast the
in()
function, and also put thevectors
in its own array with[]
to broadcast against - also learned innerjoin(), intersect(), union() to compare and merge two data-frames
-
fixing Julia Versions, after installing a new version of julia
-
Pkg.add("IJulia"); Pkg.build("IJulia")
this will re-register the new kernal which Jupyter notebooks etc and anything that requires the kernal. -
Showed cloudstor.aarnet.edu.au to transfer < 1TB of data internationally (upload via webpage, or use curl https://www.solved.tips/aarnet-cloudstor-using-curl-command-line/ )
-
Played around with visualising PC1, PC2, PC3 across RNAseq datasets using ggplot - use facet_wrap(~GROUP) to check how groups look
Attendants: 7
- turn your
for
loops intolapply()
instead - Install Rtools for Windows to install
{Rcpp}
and other packages properly on Windows. - Wrote a tiny Python script that parses an Ensembl gff3 to pull out functional annotation for mRNAs
- Discussed modeling of turtle movement - showed the usefulness of tidymodels and the upcoming tidymodels book https://www.tmwr.org/
- Showed the Profiler in RStudio for long-running code
- Introduced Julia to people who need to run many computationally-intensive models