Want to only view only a subset of samples within a dataset?
Want to define your own groups for KM plots and other analyses?
Here are the steps taken in the walk-through:
- Select the TCGA BRCA cohort
- Select TP53 expression and mutation
- Find samples where column B (expression) is less than 10 (i.e. have relatively low expression)
- Find samples that have a nonsense or frame_shift mutation for TP53
- Filter down to just these samples
- Clear this filter
- Subgroup. Make a new column where the highlighted samples are marked as 'true' and those that are not are marked as 'false'.
- Examine new subgroups in Chart mode
You can use boolean operators, search operators and other punctuation to get more specific search results.
Searching Sample IDs
Whether the sample ID column is in view or not, it is always possible to enter a sample ID to find that one sample of interest. An example:
Try using 'alt-click' to freeze the tooltip for copying a sample ID. Paste into highlight feature to keep track of where that sample lies as you move columns around, changing the sort order.
Searching a specific column
To make it easy to search a specific column, we use shorthand to annotate the first column as 'A:', the second as 'B:', etc. An example is
This will search ONLY the first column for the word 'YES'. Note that we will retain your original search if you move the columns around.
Searching categorical features
Our search is 'contains' search, meaning the term you enter can be at the beginning, end or in the middle of a matched term. Our search is case-independent. An example is
will match 'Stage IIIA' and 'Stage IIA'. To specify a specific string, use quotes
Searching continuous features
You can specify a certain column and mathematical expression such as
which will highlight all values greater than 2 in the first column. We support the following operators
- = (equal)
- >= (less than or equal)
- >= (greater than or equal)
- < (less than)
- > (greater than)
- != (not equal)
Using boolean operators and grouping
You can enter multiple search terms and we will match all of them with an implicit 'AND'. We also support 'OR'.
Use parentheses to group search terms. For example:
"Stage II" (B:Negative OR C:Negative)
will search for samples that match 'Stage II' in any column and are 'Negative' for either the second or third column.
Get rid of the grey lines - search, then filter
To get rid of all the nulls in all the columns
To get rid of nulls in a specific column
Highlights all samples with mutations with the protein change 'V600E'
Highlights all samples that have a column value of "male". Useful since the search for 'male' will return 'female' as well since male is contained within female.
D:frame OR nonsense
Highlights all samples where the 4th column has a value of 'frame' or 'nonsense'. Useful for finding high impact mutations.
Highlight samples with "full" data, no nulls in any of the column. To get rid of all the nulls, then click on the filter icon.
Search and Filter
After search, do you want to just view those samples?
How to filter samples