Quality Control
Ensuring a balanced dataset is crucial in cell tracking and similar biological analyses. This means that each biological repeat should carry equal weight in the analysis. A balanced dataset is essential for:
- Capturing True Biological Variation: Biological repeats are vital for capturing the inherent variability in biological systems. Equal weighting ensures accurate representation.
- Reducing Sampling Bias: Balancing the dataset helps avoid overemphasizing characteristics from any single repeat, which might not represent the broader biological context.
Note: If your data is imbalanced, consider balancing it when plotting your track metrics to prevent skewing your results.
- The assessment of the dataset balance is saved as a PDF file,
Track_Counts_Histogram.pdfstored in theResults_Folder/QC.
This section provides tools for computing and visualizing similarities between different FOVs and biological repeats based on selected track parameters using hierarchical clustering and dendrograms.
- Consistency Among Conditions: Ensuring FOVs from the same condition are more similar to each other than to those from different conditions.
- Reproducibility of Repeats: Confirming that biological repeats yield consistent results.
- Identifying Outliers: Spotting potential outliers or anomalies in the dataset.
- Assessing Experiment Consistency: Ensuring overall consistency and reproducibility.
- Track Parameters Selection: Select track parameters for similarity calculations.
- Similarity Metric: Choose a similarity metric from a dropdown list.
- Linkage Method: Select the method for calculating distances between clusters.
- Visualization: Click on "Select the track parameters and visualize similarity" to display dendrograms.
Hierarchical clustering in CellTracksColab visually suggests the presence of outliers and patterns within the data. However, relying solely on this technique might not provide the rigorous statistical analysis necessary for excluding repeats of field of view from the analysis. This approach helps users identify potential issues with specific fields of view or biological repeats, encouraging experimenters to consider their data differently to ensure accuracy. To assess the robustness of the analysis, users can vary the clustering and linkage methods to observe differences in the dendrograms produced.
- The hierarchical clustering results and dendrograms are saved as a single PDF file,
Dendrogram_Similarities.pdfinResults_Folder/QC. - This includes individual FOV similarity dendrograms and aggregated data dendrograms.
