Welcome to dpcReport

For in-depth description of the dpcReport, please refer to the About panel.

Upload data

Upload your data using the button below or analyze the preloaded data.

Accepted data formats:

  • raw data - comma-separated .csv file (for array digital PCR and droplet digital PCR).
  • QX100 - data from QX100 Droplet Digital PCR System (Life Technologies).
  • BioMark (Detailed Table Results) - BioMark (Fluidigm).
  • BioMark (Summary Table Results) - BioMark (Fluidigm).
  • amplification data: compressed (.zip) directory with amplification data from QX series.

See also exemplary data files.

Accepted file formats:

  • .csv,
  • .xlsx,
  • .xls,
  • .zip (only amplitude data).

Edit your data

Edit some properties of the input data using the menu below. You may change only change the experiment name and technical replicate ID.



Explanation

  1. Name of the run (NameOfExperiment.ReplicateID).
  2. Name of the experiment.
  3. ID of replicate.
  4. Method used to compute statistics.
  5. Mean number of template molecules per partition.
  6. Lower border of the confidence interval for the mean number of template molecules per partition.
  7. Upper border of the confidence interval for the mean number of template molecules per partition.
  8. Number of template molecules.
  9. Lower border of the confidence interval for the number of template molecules.
  10. Upper border of the confidence interval for the number of template molecules.
  11. Number of positive partitions.
  12. Total number of partitions.

Summary charts of digital PCR data

Boxplot chart

Calculate confidence intervals for boxplot.


The box-and-whisker plot above describes the relationship between the experiment and the λ value. Points belonging to the same experiment and with the same λ value may be overplotted. In this case, consult the replicate/experiment scatterchart below.

Scatter chart

Compare digital PCR experiments

Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

For example, if a p-value has value between 0 and 0.001, it has significance code ***.

The mean number of template molecules per partition

The table below contains the mean number of template molecules per partition (λ) and its confidence intervals. The confidence intervals were adjusted to assure stable 0.95 simultaneous coverage probability, which offers more reliable comparison of technical repeats.

The automatically assigned group is represented by a single number (in case of the intermediate groups, a combination of numbers separated by dots). Groups contain experiments whose mean number of template molecules per partition are not significantly different.

The chart above shows to which groups runs were assigned using the statistical test. Groups are marked using the number on a right side of a confidence interval.

Distribution of positive partitions

Check the distribution of positive partitions in the selected digital PCR run.

Moments of Poisson distribution

Table below represents four basic moments of Poisson distribution. Empirical moments are calculated using the sample data. Theoretical moments are computed using the estimated lambda value nad following relationships:

  • Mean: λ
  • Variance: λ
  • Skewness λ0.5
  • Kurtosis: λ-1

Distribution of basic digital PCR parameters

The chart below describes the distribution of chosen dPCR parameter (k or λ).

Parameters chosen below (confidence intervals level, method and so on) will be used in report.

Save chart (.svg)

There are several methods of calculating confidence intervals for binomial data (e.g., a status of partitions):

  • Agresti-Coull.
  • Asymptotic.
  • Bayesian inference.
  • Cloglog parameterization.
  • Logit parameterization.
  • Pearson-Klopper.
  • Probit parameterization.
  • Proportion test.
  • Wilson.

The Wilson confidence intervals are the most beneficial for dPCR experiment analysis, because they do not oscillate when the λ is close to 0. For a deeper analysis, please refer to Brown, 2001 (full references list in the About panel).

Download report

Use menu below to choose elements of the report and download it using 'Download report' button.

Each element of the report has an equivalent of a panel in the web server. Moreover, parameters set in the web server are carried to the report.

R code

Since dpcReport is written in R language using functions from dpcR package (see About panel), the analysis may be also conducted outside of the web server using the R environment. In this case check the box “R code used in the report generation”. The report will contain ready to use code, which reproduces selected steps of the analysis performed in dpcReport. Any parameter adjusted by the user in the web server will be accordingly used in the code.

Choose elements of the report

Be patient. The generation of the report may take few minutes.

Save report Save input data (.csv)

About

The dpcReport web-server is a part of the dpcR R package. All functionalities of the web-server are based on functions included in the package.

HTML5 Icon

dpcReport is a part of pcRuniveRsum.

Structure of dpcReport

dpcReport consists of seven panels which can be chosen using the navigation bar above and are described below.

Tables allow multiconditional filtering as well as sorting. The content of each table may be downloaded or printed using the menu bar at the top of the table.

Most of the figures in GUI are interactive. Static versions may be downloaded separately as .svg files.

  1. Input file: Responsible for importing data into GUI. In case if data is not properly read, make sure that you have chosen a proper data format. In this panel you can also change name of the experiments, replicates and assays in loaded data. Your original file will be NOT modified in this manner. The modified file in a dpcR exchange format may be downloaded using Save report panel. The volume of the droplet and its uncertainty is already predifined for most of the popular dPCR systems, but it may be also altered here. Since data formats for dPCR systems can change very rapidly, we provide users of the web server with exemplary data files. If a specific user-provided data set is not working with a specific data set, please reformat it accordin to exemplary files or use REDF, broader described in the dpcR manual.

  2. Data summary: This panel contains two subpanels. Summary table is a tabular summary of dPCR experiments in the imported data. Summary charts present the summary in the graphical form: as a boxplot and a scatterchart. The summary consists of λ values computed using Bhat's (Bhat et al. 2009) and Dube's (Dube et al. 2008) methods as well as the concentration of the template in the sample.

  3. Comparison of runs: Compares λ values of runs using Multiple Ratio Test (Burdukiewicz et. al., 2016). The first table contains results of pairwise comparisions between runs. Second table assignes runs to the groups (Piepho, 2004). The figure presents values of λ for individual runs and their groups.

  4. Advanced analysis: Contains specialized analytics tools.

    • Array spatial analysis - analyses of spatial randomness of positive partitions distributed over the dPCR array. Presents arrays in graphical form. Allows regional analysis of the array.
    • Probability distribution - studies distribution of dPCR data as well as the distribution of the λ.
  5. Save report: Interactively builds report from the conducted analysis.

  6. About: Basic information about the GUI.

  7. Quit: Press the button to quit the GUI.

References

The application was created using the methods derived from following publications:

[1] L. D. Brown, T. T. Cai and A. DasGupta. “Interval Estimation for a Binomial Proportion”. In: Statist. Sci. 16.2 (maj. 2001), pp. 101-133. DOI: 10.1214/ss/1009213286. .

[2] H. Piepho. “An Algorithm for a Letter-Based Representation of All-Pairwise Comparisons”. In: Journal of Computational and Graphical Statistics 13.2 (2004), pp. 456-466. ISSN: 10618600. .

[3] S. Dube, J. Qin and R. Ramakrishnan. “Mathematical analysis of copy number variation in a DNA sample using digital PCR on a nanofluidic device”. Eng. In: PloS one 3.8 (2008), p. e2876. ISSN: 1932-6203. DOI: 10.1371/journal.pone.0002876.

[4] S. Bhat, J. Herrmann, P. Armishaw, et al. “Single molecule detection in nanofluidic digital array enables accurate measurement of DNA copy number”. Eng. In: Analytical and bioanalytical chemistry 394.2 (2009), pp. 457-467. ISSN: 1618-2650. DOI: 10.1007/s00216-009-2729-5.

[5] R. M. Dorazio and M. E. Hunter. “Statistical Models for the Analysis and Design of Digital Polymerase Chain Reaction (dPCR) Experiments”. In: Analytical Chemistry 87.21 (2015), pp. 10886-10893. ISSN: 0003-2700. DOI: 10.1021/acs.analchem.5b02429. (visited on 02/29/2016).

[6] S. Rödiger, M. Burdukiewicz, K. Blagodatskikh, et al. “R as an Environment for the Reproducible Analysis of DNA Amplification Experiments”. In: The R Journal 7.2 (2015), pp. 127-150.

[7] M. Burdukiewicz, S. Rödiger, P. Sobczyk, et al. “Methods of comparing digital PCR experiments”. Eng. In: Biomolecular Detection and Quantification 28.9 (2016), pp. 14-19. ISSN: 2214-7535. DOI: 10.1016/j.bdq.2016.06.004.

Contact us

Michal Burdukiewicz

Stefan Roediger