10x Genomics
Chromium Single Cell Gene Expression
Cell Ranger4.0, printed on 11/14/2024
CRISPR Output Files
Cell Ranger outputs certain files that are specific to the CRISPR analysis. We briefly describe each file in the crispr_analysis sub-folder.
cells_per_protospacer.json
A JSON object of (key, value) pairs, where each key is a protospacer name and the corresponding value is a list of all cell barcodes in which that protospacer was detected.
{
"GFP": [
"AAACCCAAGCCGTTAT-1",
"AAACCCATCTCAAAGC-1",
"AAACGAACATTGCTTT-1",
"AAACGCTAGCCGATCC-1",
"AAAGAACAGAACAGGA-1",
...
],
"Control_1": [
"AAACCCAAGCCGTTAT-1",
"AAACGAACACCCTGAG-1",
"AAACGAACAGTTGTCA-1",
"AAACGAAGTATTCCTT-1",
"AAACGAAGTCTGTCAA-1",
...
protospacer_calls_per_cell.csv
A CSV file containing four columns: cell barcode, number of features (protospacers) detected in that cell, the names of detected feature(s), and the number of molecules (UMIs) for each feature in that cell. In the event that there are multiple features for a cell_barcode, the corresponding multiple entries for feature_call and num_umis will be separated by a | character.
$ head -3 protospacer_calls_per_cell.csv cell_barcode,num_features,feature_call,num_umis TGTCCTGCACCAGTTA-1,1,GFP,1085 CTCCGATAGACTCATC-1,1,GFP,1466
protospacer_calls_summary.csv
A CSV file containing five columns: a category (combination of protospacers or a particular property, like that of having no guide molecules detected), the number of cell barcodes belonging to that category, the percentage of cell barcodes in that category, the median UMI counts aggregated over protospacer(s) in the category (if any), and the standard deviation of UMI counts aggregated over protospacer(s) in the category (if any). In particular, we include the following 2 categories: "No guide molecules", which refers to cells where no guide molecules were detected, and "No confident call" which refers to cells where either no guide molecules were detected or the number of detected guide molecules were not judged by the protospacer-calling algorithm to be significantly different from the background.
$ head protospacer_calls_summary.csv protospacer_call,num_cells,pct_cells,median_umis,stddev_umis No guide molecules,0,0.0,None,None No confident call,538,11.2,None,None 1 protospacer expressed,4214,87.72,None,None More than 1 protospacer expressed,52,1.08,None,None GFP,2059.0,42.86,1345.0,1480.621745859537 GFP - Control_1,52.0,1.08,2015.5,2072.675627903681 Control_1,2155.0,44.86,1001.0,1815.816097039266
protospacer_umi_thresholds_csv.csv
A CSV file containing two columns: the protospacers used in the experiment, and the minimum number of molecules (umi_count) a cell needs to possess in order to be classified as expressing that protospacer.
$ head -3 protospacer_umi_thresholds_csv.csv protospacer, umi_threshold GFP,120 Control_1,79
protospacer_umi_thresholds_json.json
The same as the file above, only in JSON format.
Output Files That Rely on Differential Expression Analysis
The output files described below are only produced if the Feature Reference file includes at least 1 Non-Targeting guide RNA. These output files rely on differential expression analysis, which requires the presence of control cells (those that receive only Non-Targeting guide RNAs). For more information on Non-Targeting guide RNAs, please see CRISPR Algorithms Overview.
perturbation_efficiencies_by_feature.csv
This CSV file summarizes the effectiveness of the guide RNAs in perturbing their targets. The cells are grouped "by feature" i.e. combinations of guides present. The following columns are present:
| Column | Brief Description |
|---|---|
Perturbation |
Combinations of guide RNAs present. |
Target Guide |
The guide RNA whose target gene is being considered for the perturbation efficiency calculation. |
Log2 Fold Change |
The log2 fold change in expression of the target gene, relative to control cells. |
p Value |
p-value for the hypothesis test. |
Log2 Fold Change Lower Bound |
Lower confidence bound (5th percentile) for Log2 Fold Change. |
Log2 Fold Change Upper Bound |
Upper confidence bound (95th percentile) for Log2 Fold Change. |
Cells with Perturbation |
Number of cells with this perturbation i.e. this combination of guide RNAs. |
Mean UMI Count Among Cells with Perturbation |
Mean UMI count for the target gene in question, among cells with this combination of guide RNAs. |
Cells with Non-Targeting Guides |
Number of control cells i.e cells with only non-targeting guide RNAs. |
Mean UMI Count Among Cells with Non-Targeting Guides |
Mean UMI count for the target gene in question, among control cells. |
An example is provided below:
$ head -3 perturbation_efficiencies_by_feature.csv Perturbation, Target Guide, Log2 Fold Change, p Value, Log2 Fold Change Lower Bound, Log2 Fold Change Upper Bound, Cells with Perturbation, Mean UMI Count Among Cells with Perturbation, Cells with Non-Targeting Guides, Mean UMI Count Among Cells with Non-Targeting Guides Guide1|Guide2 Guide1 -2.487458049, 0.684604341, -3.8938847, -1.591207405, 37, 0.189189189, 3760, 1.199734043 Guide1|Guide2 Guide2 1.46887148, 0.273153378, -0.206506878, 2.354657695, 37, 0.054054054, 3760, 0.028723404
perturbation_efficiencies_by_target.csv
This CSV file is very similar to perturbation_efficiencies_by_feature.csv, with the only difference being that cells are grouped "by target" i.e. combinations of genes that their guide RNAs are targeting.
$ head -4 perturbation_efficiencies_by_target.csv Perturbation, Target Gene, Log2 Fold Change, p Value, Log2 Fold Change Lower Bound, Log2 Fold Change Upper Bound, Cells with Perturbation, Mean UMI Count Among Cells with Perturbation, Cells with Non-Targeting Guides, Mean UMI Count Among Cells with Non-Targeting Guides Gene1, Gene1, -3.742386941, 5.98E-07, -4.869624021, -2.7417221, 942, 0.001061571, 3760, 0.028723404 Gene2|Gene1, Gene2, -3.006368897, 0.666629078, -3.080612608, -2.920975978, 34, 0, 3760, 0.23643617 Gene2|Gene1, Gene1, 0.023108305, 1, -0.197392057, 0.283497989, 34, 0, 3760, 0.028723404
perturbation_effects_by_feature/top_perturbed_genes.csv
This CSV file lists the top 10 genes from the reference transcriptome that showed the greatest increase or decrease in expression as a result of each perturbation detected in the experiment. For each perturbation, the results are presented in 4 columns; the results for multiple perturbations are juxtaposed together. These columns are: the (human-readable) name of the gene, the Gene ID, the log2 fold change in expression, and the adjusted p-value (corrected for multiple comparisons).
perturbation_effects_by_feature/transcriptome_analysis.csv
This CSV file lists the effects of each perturbation on each gene in the reference transcriptome. Each row is specific to a gene: the first column specifies the Gene ID and the second column specifies the human-readable gene name. The subsequent columns summarize how the expression of each gene changes under each perturbation (combinations of guide RNAs). The results from each perturbation are summarized by 3 columns: mean UMI counts for that gene under that perturbation, log2 fold change relative to the expression in control cells, and adjusted p-value for this comparison. The results from different perturbations are juxtaposed with each other.
perturbation_effects_by_target/top_perturbed_genes.csv
This CSV file is very similar to perturbation_effects_by_feature/top_perturbed_genes.csv, with the only difference being that cells are grouped "by target" i.e. combinations of genes that their guide RNAs are targeting.
perturbation_effects_by_target/transcriptome_analysis.csv
This CSV file is very similar to perturbation_effects_by_feature/transcriptome_analysis.csv, with the only difference being that cells are grouped "by target" i.e. combinations of genes that their guide RNAs are targeting.