pySCENIC cannot reproduce AUCell + tSNE projection from a tutorial

[ninhleba]

Discussed in https://github.com/aertslab/SCENIC/discussions/386

^{Originally posted by **ninhleba** March 20, 2023} Hi, I've been testing pySCENIC by following this tutorial: http://htmlpreview.github.io/?https://github.com/aertslab/SCENICprotocol/blob/master/notebooks/SCENIC%20Protocol%20-%20Case%20study%20-%20Cancer%20data%20sets.html. The dataset I tried was the first one (Accession ID: GSE115978, Cancer type: SKCM). Everything looked pretty compatible until the AUCell + tSNE projection, which didn't seem to be able to cluster cells by cell type at all, in contrast with the corresponding plot in the tutorial.  Previous steps such as grn, ctx and aucell ran smoothly, and the clustermap based on binarized AUCell matrix I obtained also suggests tSNE computed from AUCell scores should be able to cluster out at least half the number of cell types.    I understand the nature of the algorithm makes the outputs of different runs vary slightly from one another but I don't think they should by this much. I don't think the error stems from tSNE because the PCA + tSNE projection looks like it's supposed to.  I would really appreciate it if someone could share their thoughts on this. Here are the databases that I used: - Human TFs: https://github.com/aertslab/pySCENIC/blob/master/resources/lambert2018.txt - Ranking databases: https://resources.aertslab.org/cistarget/databases/homo_sapiens/hg19/refseq_r45/mc9nr/gene_based/ + hg19-500bp-upstream-10species.mc9nr.genes_vs_motifs.rankings.feather + hg19-tss-centered-5kb-10species.mc9nr.genes_vs_motifs.rankings.feather + hg19-tss-centered-10kb-10species.mc9nr.genes_vs_motifs.rankings.feather - Motifs annotation: https://resources.aertslab.org/cistarget/motif2tf/motifs-v9-nr.hgnc-m0.001-o0.0.tbl Here is my session info: ``` ----- anndata 0.8.0 cytoolz 0.12.1 matplotlib 3.7.1 numpy 1.23.5 pandas 1.5.3 pyscenic 0.12.1 scanpy 1.9.3 seaborn 0.12.2 session_info 1.0.0 ----- PIL 9.4.0 appnope 0.1.2 asttokens NA attr 22.1.0 backcall 0.2.0 boltons NA cffi 1.15.1 cloudpickle 2.2.1 comm 0.1.2 ctxcore 0.2.0 cycler 0.10.0 cython_runtime NA dask 2023.3.0 dateutil 2.8.2 debugpy 1.5.1 decorator 5.1.1 defusedxml 0.7.1 entrypoints 0.4 executing 0.8.3 frozendict 2.3.5 fsspec 2023.3.0 h5py 3.8.0 ipykernel 6.19.2 ipython_genutils 0.2.0 jedi 0.18.1 jinja2 3.1.2 joblib 1.2.0 jupyter_server 1.23.4 kiwisolver 1.4.4 llvmlite 0.39.1 loompy 3.0.7 lxml 4.9.1 markupsafe 2.1.1 matplotlib_inline 0.1.6 mpl_toolkits NA natsort 8.3.1 networkx 3.0 numba 0.56.4 numexpr 2.8.4 numpy_groupies 0.9.20 openpyxl 3.1.1 packaging 22.0 parso 0.8.3 pexpect 4.8.0 pickleshare 0.7.5 pkg_resources NA platformdirs 2.5.2 prompt_toolkit 3.0.36 psutil 5.9.0 ptyprocess 0.7.0 pure_eval 0.2.2 pyarrow 11.0.0 pycparser 2.21 pydev_ipython NA pydevconsole NA pydevd 2.6.0 pydevd_concurrency_analyser NA pydevd_file_utils NA pydevd_plugins NA pydevd_tracing NA pygments 2.11.2 pyparsing 3.0.9 pytz 2022.7 scipy 1.10.1 setuptools 65.6.3 six 1.16.0 sklearn 1.2.2 stack_data 0.2.0 statsmodels 0.13.5 tblib 1.7.0 threadpoolctl 3.1.0 tlz 0.12.1 toolz 0.12.0 tornado 6.2 tqdm 4.65.0 traitlets 5.7.1 typing_extensions NA wcwidth 0.2.5 yaml 6.0 zmq 23.2.0 zoneinfo NA ----- IPython 8.10.0 jupyter_client 7.4.9 jupyter_core 5.2.0 jupyterlab 3.5.3 notebook 6.5.2 ----- Python 3.10.9 (main, Mar 8 2023, 04:44:36) [Clang 14.0.6 ] macOS-10.16-x86_64-i386-64bit ----- Session information updated at 2023-03-20 14:01 ```

[TMBJ-lab]

Do you solve this problem now?