spec2vec

https://spec2vec.readthedocs.io/
Admin
Word2Vec-based similarity measure for mass spectrometry data that converts spectrum peaks into...
Tokens:1,240
Snippets:14
Trust Score:9.6
Update:4 months ago
Show doc for...
Tokens:
Raw
### Install spec2vec using pip

Source: https://spec2vec.readthedocs.io

Alternative installation using pip.

```bash
pip install spec2vec
```

--------------------------------

### Install spec2vec using conda

Source: https://spec2vec.readthedocs.io

Recommended installation using Anaconda Cloud.

```bash
conda create --name spec2vec python=3.8
conda activate spec2vec
conda install --channel nlesc --channel bioconda --channel conda-forge spec2vec
```

--------------------------------

### Train a word2vec model

Source: https://spec2vec.readthedocs.io

Code example to process a dataset of reference spectra and train a word2vec model from scratch.

```python
import os
from matchms.filtering import add_losses
from matchms.filtering import add_parent_mass
from matchms.filtering import default_filters
from matchms.filtering import normalize_intensities
from matchms.filtering import reduce_to_number_of_peaks
from matchms.filtering import require_minimum_number_of_peaks
from matchms.filtering import select_by_mz
from matchms.importing import load_from_mgf
from spec2vec import SpectrumDocument
from spec2vec.model_building import train_new_word2vec_model

def spectrum_processing(s):
    """This is how one would typically design a desired pre- and post-
    processing pipeline."""
    s = default_filters(s)
    s = add_parent_mass(s)
    s = normalize_intensities(s)
    s = reduce_to_number_of_peaks(s, n_required=10, ratio_desired=0.5, n_max=500)
    s = select_by_mz(s, mz_from=0, mz_to=1000)
    s = add_losses(s, loss_mz_from=10.0, loss_mz_to=200.0)
    s = require_minimum_number_of_peaks(s, n_required=10)
    return s

# Load data from MGF file and apply filters
spectrums = [spectrum_processing(s) for s in load_from_mgf("reference_spectrums.mgf")]

# Omit spectrums that didn't qualify for analysis
spectrums = [s for s in spectrums if s is not None]

# Create spectrum documents
reference_documents = [SpectrumDocument(s, n_decimals=2) for s in spectrums]

model_file = "references.model"
model = train_new_word2vec_model(reference_documents, model_file, iterations=[10, 20, 30],
                                 workers=2, progress_logger=True)

```

--------------------------------

### Derive spec2vec similarity scores

Source: https://spec2vec.readthedocs.io

Code example to calculate similarities between mass spectrums using a trained word2vec model.

```python
import gensim
from matchms import calculate_scores
from spec2vec import Spec2Vec

# query_spectrums loaded from files using https://matchms.readthedocs.io/en/latest/api/matchms.importing.load_from_mgf.html
query_spectrums = [spectrum_processing(s) for s in load_from_mgf("query_spectrums.mgf")]

# Omit spectrums that didn't qualify for analysis
query_spectrums = [s for s in query_spectrums if s is not None]

# Import pre-trained word2vec model (see code example above)
model_file = "references.model"
model = gensim.models.Word2Vec.load(model_file)

# Define similarity_function
spec2vec = Spec2Vec(model=model, intensity_weighting_power=0.5,
                    allowed_missing_percentage=5.0)

```

--------------------------------

### Calculate scores on all combinations of reference spectrums and queries

Source: https://spec2vec.readthedocs.io

This code snippet demonstrates how to calculate scores between reference spectrums and query spectrums using the spec2vec library, and then find and print the top 10 best matches for a specific query.

```python
# Calculate scores on all combinations of reference spectrums and queries
scores = calculate_scores(reference_documents, query_spectrums, spec2vec)

# Find the highest scores for a query spectrum of interest
best_matches = scores.scores_by_query(query_documents[0], sort=True)[:10]

# Return highest scores
print([x[1] for x in best_matches])
```
Tokens:
Raw