curvefit

A command line tool for nonlinear regression analysis using the Levenberg-Marquardt algorithm.

Supported Models

- Boltzmann sigmoid:

  y = min + ((max - min)/(1 + exp((v50 - x)/slope)))

- Exponential decay:

  y = a + b * exp(-lambda * x)

- Gaussian function:

  y = a + b * exp(-mu^2/sigma^2)

- Hill equation (allosteric kinetics)

  y = delta / (1 + (ic50 / x)^hill)

- Dose response (ic50):

  y = 1 - (delta / (1 + (ic50 / x)^hill))

- Michaelis-Menten:

  y = Vmax * x / (Km + x)

- Modified sine wave

  y = a * sin(pi * (x - b) / c)

The program finds values for the parameters of the model function (e.g. a, b, c, d) that minimizes the difference between input and fitted values of y for all values of x.

Requirements

GNU Scientific Library

Usage

The command to run the program is

./curvefit -f datafile -m <model> [-a <float> -b <float> -c <float> -d <float>]

Required Arguments

-f Path to input data. Must be a tab delimited text file of (x, y) value pairs. No other data should be included in the file.

-m Desired model. Valid options are:

    boltzmann   -   Boltzmann sigmoid
    expdecay    -   Exponential decay
    gaussian    -   Gaussian
    hill        -   Hill equation
    ic50        -   Dose Response
    mm          -   Michaelis-Menten
    modsin      -   Sine wave

Optional Arguments

-abcd Specify initial values for the parameters to fit. Default values are 1.0.

    flag    boltzmann expdecay  gaussian  hill     ic50     mm      modsin
    ----------------------------------------------------------------------
    -a      min       a         a         delta    delta    Vmax    a
    -b      max       b         b         ic50     ic50     Km      b
    -c      v50       lambda    mu        hill     hill     N/A     c
    -d      slope     N/A       sigma     N/A      N/A      N/A     N/A