diff --git a/src/components/data-processing/CellSizeDistribution/CalculationConfig.jsx b/src/components/data-processing/CellSizeDistribution/CalculationConfig.jsx index 50826b0c8c..45a9624ca7 100644 --- a/src/components/data-processing/CellSizeDistribution/CalculationConfig.jsx +++ b/src/components/data-processing/CellSizeDistribution/CalculationConfig.jsx @@ -26,7 +26,7 @@ const CellSizeDistributionConfig = (props) => { - + { {' '} function - . This - distinguishes between droplets containing cells and ambient RNA. The FDR range is + . This distinguishes between real cells and background (ambient RNA). The FDR range is [0-1]. The default FDR value is 0.01, where only barcodes with FDR < 0.01 are retained. diff --git a/src/pages/experiments/[experimentId]/data-processing/index.jsx b/src/pages/experiments/[experimentId]/data-processing/index.jsx index 8f9789ad66..a1ae78cc15 100644 --- a/src/pages/experiments/[experimentId]/data-processing/index.jsx +++ b/src/pages/experiments/[experimentId]/data-processing/index.jsx @@ -175,7 +175,7 @@ const DataProcessingPage = ({ experimentId, experimentData }) => { { key: 'classifier', name: getUserFriendlyQCStepName('classifier'), - description: 'The Classifier filter is based on the ‘emptyDrops’ method which distinguishes between droplets containing cells and ambient RNA. Droplets are filtered based on the False Discovery Rate (FDR) value - the red line on the density plot. In the knee plot, the ‘mixed’ population shown in grey contains some cells that are filtered out and some that remain and can be filtered further in the next filter.', + description: 'The Classifier filter is based on the "emptyDrops" method which distinguishes between real cells and background (ambient RNA). Barcodes are filtered based on the False Discovery Rate (FDR) value - the red line on the density plot. In the knee plot, the "mixed" population shown in grey contains some cells that are filtered out and some that remain and can be filtered further in the next filter.', multiSample: true, render: (key) => ( { { key: 'cellSizeDistribution', name: getUserFriendlyQCStepName('cellSizeDistribution'), - description: 'The number of unique molecular identifiers (#UMIs) per cell distinguishes real cells (high #UMIs per cell) from empty droplets (low #UMIs per cell). This filter is used to detect empty droplets and fine-tunes the Classifier filter. In some datasets this filter might be used instead of the Classifier filter.', + description: 'The number of unique molecular identifiers (#UMIs) per cell distinguishes real cells (high #UMIs per cell) from background (low #UMIs per cell). This filter is used to detect background and fine-tunes the Classifier filter. In some datasets this filter might be used instead of the Classifier filter.', multiSample: true, render: (key) => ( { { key: 'mitochondrialContent', name: getUserFriendlyQCStepName('mitochondrialContent'), - description: 'A high percentage of mitochondrial reads is an indicator of cell death. UMIs mapped to mitochondrial genes are calculated as a percentage of total UMIs. The percentage of mitochondrial reads depends on the cell type. The typical cut-off range is 10-50%, with the default cut-off set to 3 median absolute deviations above the median.', + description: 'A high percentage of mitochondrial reads is an indicator of cell death. UMIs mapped to mitochondrial genes are calculated as a percentage of total UMIs. The percentage of mitochondrial reads depends on the cell type. The typical cut-off range is 5-30%, with the default cut-off set to 3 median absolute deviations above the median.', multiSample: true, render: (key) => ( { { key: 'numGenesVsNumUmis', name: getUserFriendlyQCStepName('numGenesVsNumUmis'), - description: 'The number of expressed genes per cell and number of UMIs per cell is expected to have a linear relationship. This filter is used to exclude outliers (e.g. many UMIs originating from only a few genes).', + description: 'The number of expressed genes per cell and number of UMIs per cell is expected to have a linear relationship, until the maximum number of genes is reached and the curve tends to plateau. This filter is used to exclude outliers (e.g. many UMIs originating from only a few genes).', multiSample: true, render: (key) => ( { name: getUserFriendlyQCStepName('doubletScores'), description: ( - Droplets may contain more than one cell. + A single barcode might correspond to more than one cell. In such cases, it is not possible to distinguish which reads came from which cell. - Such “cells” cause problems in the downstream analysis as they appear as an intermediate type. - “Cells” with a high probability of being a doublet should be excluded. - The probability of being a doublet is calculated using ‘scDblFinder’. + Such barcodes cause problems in the downstream analysis as they appear as an intermediate type. + Barcodes with a high probability of being a doublet should be excluded. + The probability of being a doublet is calculated using "scDblFinder". For each sample, the default threshold tries to minimize both the deviation in the expected number of doublets and the error of a trained classifier. For more details see {' '} diff --git a/src/redux/reducers/componentConfig/initialState.js b/src/redux/reducers/componentConfig/initialState.js index a3ff1e62ca..d54a48acb2 100644 --- a/src/redux/reducers/componentConfig/initialState.js +++ b/src/redux/reducers/componentConfig/initialState.js @@ -648,8 +648,8 @@ const classifierKneePlot = { }, axes: { ...axesBaseState, - xAxisText: 'Droplet Rank', - yAxisText: 'Droplet #UMIs', + xAxisText: 'Barcode Rank', + yAxisText: 'Barcode #UMIs', }, axesRanges: { ...axesRangesBaseState,