Help
This page covers usage of this web application. For general information about MuSyC and frequently asked questions, see the About page.
Input data format
Users' data is uploaded in comma-separated value (CSV) format, using the Unicode UTF-8 encoding (the default across most software). The required columns and their contents are shown in the table below. Combinations are grouped based on drug1, drug2, and sample columns. Thus, multiple combinations can be uploaded in the same file and will be processed separately. The drug1.units and drug2.units are arbitrary but must be consistent within each combination.
| Column Name | Data Type | Description |
|---|---|---|
| expt.date | yyyy-mm-dd | Experiment date |
| drug1.conc | Float | Drug 1 concentration |
| drug2.conc | Float | Drug 2 concentration |
| effect | Float | Effect value (e.g., % viability, DIP rate) |
| sample | String | Sample name (e.g., name of cell line) |
| drug1 | String | Drug 1 name |
| drug2 | String | Drug 2 name |
| drug1.units | String | Drug 1 units (e.g., nM) |
| drug2.units | String | Drug 2 units (e.g., nM) |
| effect.95ci | Float (optional) | 95% confidence interval for effect value |
Create an account
Go to Create an account. Please enter your email address and a choice of password. Your email address will be verified by sending an email with a clickable link. Click on this link, and you will then be able to log in to the site with your credentials.Create a dataset
After logging in, click the Create dataset button. Fill out the form with each of the required fields:- Name: A name to refer to the dataset
- File: Select a CSV file from your computer, formatted as described above
-
Orientation: Whether Emax > E0, or Emax < E0:
Above: Dose-response curves showing Emax > E0 (left) and Emax < E0 (right)
- Metric name: The name of the drug effect (e.g., "Percent effect" or "DIP rate")
-
Effect constraint: Whether to fix or apply bounds to the E0
and Emax values.
- Unconstrained: Emax and E0 are determined empirically
- Fixed: Emax and E0 are specified manually. This implies no synergistic efficacy, i.e. beta=0. You can specify one fixed value and leave the other one unconstrained (blank) if desired.
- Upper/lower bounds: Specify range bounds for Emax and E0. You can provide lower, upper, or both bounds, for either or both of Emax or E0. Leave blank for unconstrained.
After completing the form, click the "Create dataset" button on this page. The file will be uploaded and fitting tasks created for each drug combination. Depending on the dataset size and demand, this process make take a few seconds to several minutes. Do not navigate away from the page while the upload is in progress.
The Dataset Page
After a dataset upload is complete, you will see the dataset page. You can also return to the dataset page at any time by logging in and by clicking the dataset's name on the home page.
Each combination experiment is sent to a queue for processing. The processing step runs the MuSyC algorithm to fit the dose-response surface, and returns the relevant fitting parameters. For small datasets, this process typically only takes a few minutes, but this will vary depending on dataset size and server demand.
When the upload is complete, the web browser will redirect to a page showing the dataset name and a list of the combination experiments in a table. A progress bar indicates whether there are still fitting tasks queued, in progress, or completed. At the bottom of the page, there is a link to download the dataset's fitting parameters as a CSV file.
Each drug combination in the table shows the fitting algorithm's status (e.g., queued, started, success, failed). For successful tasks, clicking on the word "SUCCESS" will show the parameters for that combination, along with an interactive dose-response surface plot, which can be zoomed, panned, rotated etc. in the web browser. For tasks marked as "FAILED", clicking on that word will show more details about the error (e.g., if there was a data validation issue that the user should correct). There is also a link at the bottom of the task result page to download that single combination's parameters as a CSV file.
Parameter CSV file
The following table gives a description of each of the fields in the parameter CSV files.| Parameter | Description |
|---|---|
| sample | Sample name |
| drug1_name | Name of compound 1 (d1) |
| drug2_name | Name of compound 2 (d2) |
| expt | Name of experiment |
| batch | Name of the batch (if applicable) |
| task_status | Did the task succeed? |
| converge_mc_nlls | Did the Monte Carlo sampling converge? (1=True) |
| beta | The % increase of in the effect of the combination over the most efficacious single agent (based on fitted Emax). Beta = (min(E1,E2)-E3)/(E0-min(E1,E2)) |
| beta_ci | The 95% confidence interval for beta based on Monte Carlo sampling |
| beta_obs | The observed % increase of in the effect of the combination over the most efficacious single agent (based value of fit at max tested conc.). Beta_obs = (min(E1_obs,E2_obs)-E3_obs)/(E0-min(E1_obs,E2_obs)) |
| beta_obs_ci | The 95% confidence interval for beta_obs based on Monte Carlo sampling |
| log_alpha12 | Log of synergistic potency (drug1’s affect on potency of drug2). Values < 0 are antagonistically potent. Values >0 are synergistically potent. Alpha1 is quantifies the shift in the EC50 of drug 2 at saturating conc of d1. |
| log_alpha12_ci | Uncertainty in log_alpha1 |
| log_alpha21 | Log of synergistic potency (drug2’s affect on potency of drug1). Values < 0 are antagonistically potent. Values >0 are synergistically potent. Alpha2 is quantifies the shift in the EC50 of drug 1 at saturating conc of d2. |
| log_alpha21_ci | Uncertainty in log_alpha2 |
| R2 | R-squared of fit |
| log_like_mc_nlls | Log likelihood of the parameter set |
| E0 | Fitted basal effect when [d1]=[d2]=0 |
| E0_ci | The 95% confidence interval for E0 |
| E1 | Fitted effect for [d1]->inf and [d2]=0 |
| E1_ci | The 95% confidence interval E1 |
| E2 | Value of fit at max tested conc. d1 and [d2]=0 |
| E2_ci | The 95% confidence interval E1_obs |
| E3 | Fitted effect for [d2]->inf and [d1]=0 |
| E3_ci | The 95% confidence interval E2 |
| E1_obs | Value of fit at max tested conc. d2 and [d1]=0 |
| E1_obs_ci | The 95% confidence interval E2_obs |
| E2_obs | Fitted effect for [d1]->inf and [d1]->inf |
| E2_obs_ci | The 95% confidence interval E3 |
| E3_obs | Value of fit at max tested conc. d1 and d2 |
| E3_obs_ci | The 95% confidence interval E3_obs |
| log_C1 | Log of the EC50 for drug 1 |
| log_C1_ci | Uncertainty in log_C1 |
| log_C2 | Log of the EC50 for drug 2 |
| log_C2_ci | Uncertainty in log_C2 |
| log_h1 | Log of the hill slope d1 |
| log_h1_ci | 95% confidence interval of log_h1 |
| log_h2 | log of hill slope d2 |
| log_h2_ci | 95% confidence interval of log_h2 |
| h1 | Hill slope d1 |
| h2 | Hill slope d2 |
| C1 | EC50 drug1 |
| C2 | EC50 drug2 |
| time_total | Time to fit |
| drug1_units | Drug 1 units |
| drug2_units | Drug 2 units |
| metric_name | name of metric of drug effect |
| fit_beta | Was Emax fixed such that beta=0? 1=Yes; 0=No |
| boundary_sampling | Was boundary sampling fit used |
| max_conc_d1 | maximum tested concentration of drug 1 |
| max_conc_d2 | maximum tested concentration of drug 2 |
| min_conc_d1 | minimum tested concentration of drug 1 |
| min_conc_d2 | minimum tested concentration of drug 2 |
| fit_method | What method was used to fit the 2D Hill equation |
| dataset_name | Name of dataset in MuSyC portal |
Please send questions, issues, and/or commercial licensing requests to: musyc@gmail.com.
Create an account Back to Home Page