Synchronized navigation of contact maps - ma-compbio/Higashi GitHub Wiki

In this documentation, all panels would be marked with Bold Italic, any components (buttons, sliders, menus) in the tool would be marked with code format. Similar to before, the file names are also marked with code format.

Synchronized navigation of single cell contact maps

A quick review of the available tools on the Embedding visualization panel:

  • Pan: when activated, the mouse drag operation would be defined as moving the coordinates.
  • Lasso select: when activated, the mouse drag operation would be defined as the lasso selection of dots.
  • Box select: when activated, the mouse drag operation would be defined as box selection (rectangular shape) of dots.
  • Zoom: when activated, the mouse scroll wheel can be used to zoom in and zoom out of the scatter plots.
  • Click select: when activated, the user can click on the dots to select the dots
  • Reset: when clicked, the scatter plot would be restored to the default scale and range.
  • Save figure: when clicked, the user can save the scatterplot.
  • Display info when hover: when activated, when the user hovers the mouse over a dot, detailed information of the corresponding cell would show.

One of the key functions of Higashi-vis is integrative navigation of single-cell contact maps and embedding vectors. Specifically, when select one or multiple dots in the scatter plot, the corresponding (pooled) single-cell contact maps would be visualized in the Contact maps visualization panel. Both raw and imputed contact maps would be visualized for making comparisons.

Selecting the cells of interest and inspect the corresponding contact maps is key to the single-cell Hi-C analysis. In Higashi-vis, we provided multiple ways of selecting cells that are user-friendly and intuitional.

Select the cell(s) by clicking on the dot(s)

The most straightforward way of selecting a cell is by clicking the corresponding dot(s) in the scatter plot. Notably, when selecting a single cell, the selected cell will be colored red in the scatter plot, the neighboring cells that help with the imputation will be colored orange, and the rest cells will be colored blue.

Select a specific cell with index

When the user has a specific cell in mind that wants to inspect, please use the cell selector slider to select the cell based on the index. The Previous and Next buttons right above the cell selector slider could be used to fine-tune the selection of a specific cell. The cell index is defined in the data.txt, which is part of the input.

Select cells with lasso selection tool

To select a group of cells, one can use the lasso selection tool. Notably, when multiple cells are selected. The distribution of the label information used as the color scheme within the selected group of cells will be highlighted in the Statistics visualization panel. When hovering over the highlighted bars, the count information would be displayed.

Select cells with box selection tool

Besides lasso selection tool, box selection tool can also be used for selecting a group of cells.

Select cells that share the same label

Higashi also allows users to select a group of cells that share the same label. Specifically, when clicking on a bar in the Statistics visualization panel, cells that have this label would all be selected.

Select a local cell neighborhood by click

For datasets with lower coverage, sometimes it is valuable to inspect contact maps at a pseudo-cell level. The local neighborhood selection function allows automatical generation of pseudo-cell. One should first define the size of the pseudo-cell with the local selection number slider. After that, by clicking a cell, its k-nearest cells in the visualized scatter plot would all be selected.

Note: this function only works when the user clicks on one single dot. Sometimes when dots are overlapping with each other, one click may select multiple cells which would not trigger the local selection function. To fix that, one can either use a smaller dot size with the scatter size slider or zoom in on the scatter plot, etc. Moreover, the k-nearest-neighbors are defined in the 2-dimensional space, so it's sensitive to the choice of the visualization method.

Select a local cell neighborhood by index

Similar to selecting a local cell neighborhood by click. The only difference is instead of clicking on one cell, here, one can select a cell with its local neighborhood by cell index with the cell selector slider.