Documentation for Noise.py

Noise

GaussianPeakNoise

GaussianPeakNoise(sigma, log_space=False)

Bases: NoPeakNoise

Adds Gaussian noise to peaks

Initialises Gaussian peak noise

Parameters:

Name	Type	Description	Default
sigma		the variance	required
log_space		whether to sample in log space	False

Source code in vimms/Noise.py

def __init__(self, sigma, log_space=False):
    """
    Initialises Gaussian peak noise

    Args:
        sigma: the variance
        log_space: whether to sample in log space
    """
    self.sigma = sigma
    self.log_space = log_space

get

get(original, ms_level)

Get peak measurement with gaussian noise applied

Parameters:

Name	Type	Description	Default
original		original value	required
ms_level		ms level	required

Returns: peak measurement with gaussian noise applied

Source code in vimms/Noise.py

def get(self, original, ms_level):
    """
    Get peak measurement with gaussian noise applied

    Args:
        original: original value
        ms_level: ms level

    Returns: peak measurement with gaussian noise applied

    """
    return trunc_normal(original, self.sigma, self.log_space)

GaussianPeakNoiseLevelSpecific

GaussianPeakNoiseLevelSpecific(sigma_level_dict, log_space=False)

Bases: NoPeakNoise

Adds ms-level specific Gaussian noise to peaks

Create a gaussian peak noise level specific

Parameters:

Name	Type	Description	Default
sigma_level_dict		key: level, value: sigma. ms_levels not in the dict will not have noise added allows noise to be added to oa single level, or to all levels with different sigma	required
log_space		whether to log or not	False

Source code in vimms/Noise.py

def __init__(self, sigma_level_dict, log_space=False):
    """
    Create a gaussian peak noise level specific

    Args:
        sigma_level_dict: key: level, value: sigma.
                          ms_levels not in the dict will not have noise added
                          allows noise to be added to oa single level, or
                          to all levels with different sigma
        log_space: whether to log or not
    """
    self.log_space = log_space
    self.sigma_level_dict = sigma_level_dict

NoPeakNoise

The base peak noise object that doesn't add any noise

get

get(original, ms_level)

Get the original value back. No noise if applied.

Parameters:

Name	Type	Description	Default
original		The original value	required
ms_level		The ms level	required

Returns: the original value (unused)

Source code in vimms/Noise.py

def get(self, original, ms_level):
    """
    Get the original value back. No noise if applied.

    Args:
        original: The original value
        ms_level: The ms level

    Returns: the original value (unused)

    """
    return original

UniformSpikeNoise

UniformSpikeNoise(density, max_val, min_val=0, min_mz=None, max_mz=None)

A class to add uniform spike noise to the data

Create a UniformSpikeNoise class Args: density: number of spike peaks per mz unit max_val: maximum value of spike min_val: minimum value of spike min_mz: maximum m/z max_mz: minimum m/z

Source code in vimms/Noise.py

def __init__(self, density, max_val, min_val=0, min_mz=None, max_mz=None):
    """
    Create a UniformSpikeNoise class
    Args:
        density: number of spike peaks per mz unit
        max_val: maximum value of spike
        min_val: minimum value of spike
        min_mz: maximum m/z
        max_mz: minimum m/z
    """
    self.density = density
    self.max_val = max_val
    self.min_val = min_val
    self.min_mz = min_mz
    self.max_mz = max_mz

trunc_normal

trunc_normal(mean, sigma, log_space)

Ensures that generators never return negative mz or intensity

Parameters:

Name	Type	Description	Default
mean		mean of gaussian distribution to sample from	required
sigma		variance of gaussian distribution to sample from	required
log_space		whether to sample in log space	required

Returns: the sampled value

Source code in vimms/Noise.py

def trunc_normal(mean, sigma, log_space):
    """
    Ensures that generators never return negative mz or intensity

    Args:
        mean: mean of gaussian distribution to sample from
        sigma: variance of gaussian distribution to sample from
        log_space: whether to sample in log space

    Returns: the sampled value

    """
    s = -1
    if not log_space:
        while s < 0:
            s = np.random.normal(mean, sigma, 1)[0]
        return s
    else:
        s = np.random.normal(np.log(mean), sigma, 1)[0]
        return np.exp(s)