SAPIEnS Database description

Aims of the SAPIEnSdb development

• Provide a set of TFs specific for a particular cell type based on digital genomic footprinting in the scATAC-seq data
• Provide a posibility to compare TF sets obtained using different preprocessing and imputation methods

List of the datasets

One can find list of available footprint-based datasets in the base at Datasets page.

Dataset Page

This page describes experiments provided for our database. One can access this page using link.

Dataset page reflects three ways to access the data:

1. Tab "Experiment list" shows the experiments with completed RGT-HINT results. We have uploaded results for the best iteration after every imputation method. Link to the experiment number (column #) allows one to get to the "Activity Score" page
2. Tab "Built-in graphs" allows one to select one of the built-in evolution graphs. An evolution graph shows transitions between cell types in the dataset.
3. Tab "Compare experiments by TF" allows one to query whether a selected imputation factor is revealed significant between two available imputation frameworks.

Activity Scores calculation

The primary aim of SAPIEnSdb is to find significant evolution parts between activity scores. Table shows the list of all available transcription factors motifs from HOCOMOCO.

Page for HSC available at link: HSC, no imputation. If one needs to find the activated transcription factors of category A in MPP branch, one should select:

1. Checkbox "A" in column "Motif name".
2. Press ">2" in column MEP.

After that one will get a list of transcription factors that have Activity Scores higher than 2.

Graph view page

Graph view page shows one a evolution tragecory between populations. One gets a list of branches that is tested in the evolution graph.

One can see a tragectory of Haematopoietic Development. One can reveal the significant transitions between two cell types by selecting an imputation framework and a TF. For example, one can choose SCALE and Boruta imputation framework and a GATA2 TF. The obtained graph shows that a differentially significant transition has found between CMP and MEP branches.

Comparison of imputation frameworks

The third feature of our database allows to find transcription factors that are differential only for specific experiment. For example, one needs to compare the scOpen and SCALE imputation methods with Boruta preprocessing.

After pressing button "Compare" one can see whether TFs are significant between each pair of cell types. For example, we see that using scOpen method reveals that GATA1 is significant between Lympho-myeloid Primed Progenitor cells and Haematopoieitic Stem Cells whether SCALE does not show significance. This page can be used to reveal significant TFs along the tragectories.