Download the CRISPR
dataset (Achilles_gene_effect.csv). This dataset facilitates the
discovery of genes that when knocked down/out effect the viability of particular cell
lines.
Morpheus can also open data from Excel, text files,
the clipboard, a URL, and Dropbox.
Select Tools > Transpose to put the genes on the rows and the cell lines on the columns.
Adjust Color SchemeSelect View > Options to edit the color scale. By default,
values in the heat map are mapped to colors
using the minimum and maximum of each row independently.
With z-scored data, a fixed scale should be used. Set the range of the fixed color scale to -2, 2.
Video
Download the cell line metadata (sample_info.csv). Select File > Open and then select "Annotate Columns" for "Open File Action". Next go to View > Options and select the Annotations tab. Select "id", "lineage", and "sample_collection_site" for column annotations.
Variance
Click Tools > Create Calculated Annotation.
Enter "variance" for
the "Annotation name".
Enter "VARIANCE()" for "Formula" and
click "OK".
A new row annotation named "variance" will appear to
the right of the heat map.
Show only the top 500 most variable genes by variance
Click Tools > Filter to open the filter dialog.
Click the "Add"
button and set the field name to "variance".
Click
"Switch to top N filter" and enter "500" for "N".
Click Tools > Hierarchical Clustering. Change "Cluster" to "Rows and
columns". Click "OK" to run the analysis.
Do haematopoietic
cell lines cluster together? Try searching the columns for "haematopoietic_and_lymphoid_tissue". Right-click on
the sample_collection_site header
and select "Highlight matching values". Hover over the sample_collection_site
values to highlight cell lines from the same site. You
can dynamically cut the dendrogram by dragging the dashed line at
the top of the dendrogram.
Hierarchical clustering
recursively merges objects based on their pair-wise distance.
Objects closest together are merged first, objects furthest apart
are merged last. The result is a tree structure, referred to as a
dendogram, where the leaf nodes represent the original items and
internal (higher) nodes represent the merges that occurred. Click here
for a more detailed description of the hierarchical clustering
algorithm and here
for a comparison of the Pearson and Spearman correlation methods.
Click View > Fit To Window to compress the heat map. Select View > 100% to return to the original heat map size.
Remove Row FilterSelect Tools > Filter and delete the variance filter to show all rows.
Remove DendrogramsRight-click on the column dendrogram and select "Delete". Do the same for the row dendrogram.
Sort RowsClick the variance row annotation header in the heat map once to sort the heat map in ascending order by the variance. Click it again to sort the heat map in descending. The third click sorts by alternating groups of ten of most similar and dissimilar items. Note that you can shift-click to sort multiple columns simultaneously.
Search Results To Top
Enter "BRAF" in the row search box.
Click
to bring the matches to the top of the heat map.
Double-click on BRAF row to sort by dependency score.
Search columns for "skin".
Do skin cell lines seem to be
more dependent on BRAF than cell lines from other lineages?
You can optionally limit your search to within a field by typing
the field name followed by a colon ":" and then the term you are
looking for.
Click Tools > Nearest Neighbors. Click "OK" to run the analysis.
A new row annotation named "Pearson correlation" will appear.
What genes correlate with BRAF?
Download the PRISM drug sensitivity data (primary_replicate_collapsed_logfold_change.csv) and select File > Open. Open the dataset in a new tab. Select Tools > Transpose to put the compounds on the rows and the cell lines on the columns. Select File > Save Dataset to save the transposed matrix and enter "prism_logfold_change.gct" for the file name. Navigate back to the gene essentiality dataset tab. Select File > Open to open the transposed drug sensitivity matrix. Change "Open File Action" to "Append rows to current dataset". Match the cell line ids in the CRISPR dataset with the cell line ids in the drug sensitivity dataset.
Repeat Nearest Neighbors With drug sensitivity dataset
Create a new tab with BRAF CRISPR dependency scores and drug sensitivity scores by searching rows for
"Source:prism_logfold_change BRAF". Ensure that all columns are selected. Click Tools > New Heat Map to
create a new tab.
Download the compound
annotations (primary_replicate_collapsed_treatment_info.csv) and select File > Open. Choose "Annotate rows" for "Open File Action".
In the new tab, search for BRAF and bring the matches to the top. Select the BRAF dependency scores. Click Tools
> Nearest Neighbors.
Click "OK" to run the analysis.
A new row annotation named "Pearson correlation" will appear.
What drugs correlate with BRAF dependency scores?
Create a new tab with the CRISPR dependency scores only. Remove the number after the gene name by selecting
Tools>Create Calculated Annotation. Enter "id_stripped" for the annotation
name and FIELD('id').substring(0, FIELD('id').indexOf('(')) for the formula.
Download the RNAi dataset (D2_Achilles_gene_dep_scores.csv) and select File>Open. Repeat the same
process
to remove the numbers after the gene name by selecting
Tools>Create Calculated Annotation. Enter "id_stripped" for the annotation
name and FIELD('id').substring(0, FIELD('id').indexOf('(')) for the formula. Select File > Save
Dataset to save the updated dataset and enter "D2_Achilles_gene_dep_scores.gct" for the file name. Navigate back
to the
CRISPR tab and select File > Open. Change "Open File
Action" to "Overlay onto
current dataset". Match the "CCLE Name" in the CRISPR dataset with "id" in the RNAi
dataset and the "id_stripped" annotation in the CRISPR dataset with "id_stripped" in the RNAi dataset. Select
View > Options to
edit the color scale. Change "shape" to "circle" and "size by" to the RNAi dataset. Set the size by minimum to 2
and the maximum to -2.
Download the CCLE Mutations dataset (CCLE_mutations.csv). Rename the file to have the suffix .maf.csv (e.g. CCLE_mutations.maf.csv) so that Morpheus will recognize the file type properly.