Processing individual samples#

We will present here both the low and high level interfaces to process data from individual samples. Before starting with these topics, we will create a simulated sample that will be used in the examples. This can be done by using the simulation module:

from tidyms2.lcms.simulation import SimulatedLCMSSampleFactory

factory_spec = {
    "config": {
        "min_signal_intensity": 1.0,
        "n_scans": 40,
        "amp_noise": 0.0,
    },
    "adducts": [
        {
            "formula": "[C54H104O6]+",
            "rt_mean": 10.0,
            "base_intensity": 1000.0,
            "n_isotopologues": 2,
        },
        {
            "formula": "[C27H40O2]+",
            "rt_mean": 20.0,
            "base_intensity": 2000.0,
            "n_isotopologues": 2,
        },
        {
            "formula": "[C24H26O12]+",
            "rt_mean": 30.0,
            "base_intensity": 3000.0,
            "n_isotopologues": 2,
        },
    ],
}

simulated_sample_factory = SimulatedLCMSSampleFactory(**factory_spec)
sample = simulated_sample_factory(id="my_sample")

Simulated samples can be used as regular sample models to access “raw” data:

from tidyms2 import MSData

data = MSData(sample)
spectrum = data.get_spectrum(10)  # get the 10th scan from the simulated data

We are going to use this simulated sample on all the examples in this page. For more details about how to create simulated samples, refer to this guide.

Low-level API#

Basic data processing from raw data can be done using low level functions available in the algorithms subpackage. For example, using the algorithms.raw module, we can create EIC by providing a list of m/z values:

from tidyms2.algorithms.raw import MakeChromatogramParameters, make_chromatograms
from tidyms2.chem import Formula

# we are using here m/z values from from two of the molecules included in the
# simulated sample.
mz1 = Formula("[C27H40O2]+").get_exact_mass()
mz2 = Formula("[C54H104O6]+").get_exact_mass()

params = MakeChromatogramParameters(mz=[mz1, mz2])
chromatograms = make_chromatograms(data, params)

If you installed the plotting optional dependencies, you can plot the chromatograms using the following function:

TODO: COMPLETE

The accumulate_spectra() combines a series of successive spectra into a single spectrum:

from tidyms2.algorithms.raw import AccumulateSpectraParameters, accumulate_spectra

params = AccumulateSpectraParameters(start_time=10.0, end_time=15.0)
accumulated_spectrum = accumulate_spectra(data, params)

The algorithms package provides several utilities to process data, such as untargeted ROI creation from raw data or peak picking routines, but you need to write the logic for moving and storing data around. The high level API provides an easy to use way to process data and orchestrates data storage between transformations.

High level API#

The high level API uses is based on the TidyMS data model. The general idea is that sample transformations are defined in a data pipeline, comprised by multiple operators that are applied one at a time to the sample data. We’ll illustrate data processing using individual operators first, and then we show how to apply multiple operators using a pipeline. It is highly recommended to read the:ref:architecture overview guide <overview> as it presents a high level yet complete view of sample data processing using pipelines and operators.

The first thing to do to process sample data through a pipeline is to create a sample data storage. We will use the :py:class`tidyms.storage.OnMemorySampleStorage` class for this:

from tidyms2.lcms import MZTrace, Peak
from tidyms2.storage import OnMemorySampleStorage

sample_data = OnMemorySampleStorage(sample, MZTrace, Peak)

Note that, besides the sample model, we also need to pass the type of data that we want to store. Sample data into two entities that are part of the TidyMS data model: ROI and features. A ROI is a data subset extracted from raw data where features may be extracted from. A feature is a ROI subregion that contains information associated with a chemical species. The sample data storage classes manage ROI/feature storage and retrieval extracted from raw data. For LC-MS data, the MZTrace models a ROI for LC-MS data and it is basically an m/z trace containing m/z and intensity values on each scan. The Peak models a chromatographic peak detected in an m/z trace. We will use the LC-MS sample operators LCTraceExtractor and LCPeakExtractor first to extract m/z traces from raw data and then detect peaks on each m/z trace:

On this page

This Page