Skip to main content
Pre-release

This release is a pre-release and may not be stable for production use.

Protspect

Protspect logo

Protein sequence spectral analysis: transform features for proteins and peptides using numeric residue-property channels. This is the original sequence featurization method, separate from the subsequent protein language-model work.

The repository is biomodsquad/protein-sequence-spectral-analysis. The PyPI distribution is Protspect, imported as protspect. The original ProSpect distribution name conflicts with an unrelated existing PyPI package. Version 0.1.0a2 is the second alpha; APIs may change.

Documentation and tutorials

Read the online documentation, covering the unified API in alpha 0.1.0a2.

The documentation source follows the MISTIC Sphinx/Furo layout: installation, matrix-based concepts, eight guided tutorials, downloadable notebooks, and a public API reference. The supplied logo and matrix illustration introduce the residue embedding → zero-padding → sequence-axis transform workflow.

Three self-contained notebooks are bundled in protspect.examples; locate them with importlib.resources.files("protspect.examples") and copy to a writable workspace. They cover matrix-to-feature calculations, four transforms and paired chains, and training-only sklearn pipelines. The full AMP MLP comparison and embedding atlas remain separate research examples with their documented inputs.

Build a local site with:

python -m pip install -e '.[docs]'
python -m docs.generate_tutorial_figures
python -m sphinx -W --keep-going -b html docs docs/_build/html

See build and validation instructions for executable snippets, notebook checks, and manual Pages publication. The documentation includes executable examples for alpha 0.1.0a2.

Install

Requires Python 3.10 or later. Install the published alpha and its dependencies:

python -m pip install Protspect==0.1.0a2
# Optional scikit-learn integration:
python -m pip install 'Protspect[ml]==0.1.0a2'

For development, install from the repository:

git clone https://github.com/biomodsquad/protein-sequence-spectral-analysis.git
cd protein-sequence-spectral-analysis
python -m venv .venv
source .venv/bin/activate
python -m pip install -e '.[dev]'

The distribution is named Protspect. The unrelated prospect package is a different project.

Alternative transforms

Fourier, circular Z, finite discrete Mellin, and discrete wavelet transforms now share transform_sequence, with a fixed-width SequenceTransformer for sklearn. See definitions and examples. Wavelets require the optional wavelets extra.

Residue representations

Protspect includes the ten Kidera factors (KF1–KF10), in addition to the default BLOSUM100 embedding. See representation documentation and citation.

from protspect import SequenceSet
from protspect.embeddings import kidera_factors

features = SequenceSet(embedding=kidera_factors(), compute_cross_spectra=False)
features.from_sequence_list(["ACDE", "ACDF"], spectrum_length=8)

AAindex1 files can also supply selected property channels:

from protspect.aaindex import AAIndex

database = AAIndex.from_file(".aaindex/aaindex1")
features = SequenceSet(
    embedding=database.embedding(["KYTJ820101", "FAUJ880103"]),
    compute_cross_spectra=False,
)

See AAindex1 loading, explicit downloads, and provenance. No database is bundled or downloaded during feature extraction.

AAindex2/3 matrices also support explicit row or column profile channels:

from protspect.aaindex import AAIndex2

matrices = AAIndex2.from_file(".aaindex/aaindex2")
embedding = matrices.embedding("HENS920102", axis="rows")  # BLOSUM62

See matrix conversion and contact-potential interpretation.

One transform API

The API supports method='fourier'|'z'|'mellin'|'wavelet' throughout. SequenceSet and ComplexSet remain primary interfaces for single and paired collections. All entry points share the same backend and feature exporter. Install Protspect[ml,wavelets]==0.1.0a2 for sklearn and wavelet support.

from protspect import SequenceSet, ComplexSet, transform_sequence

features = SequenceSet(method='fourier', compute_cross_spectra=False)
features.from_sequence_list(['ACDE', 'ACDF'], length=8,
    attributes={'name': ['peptide-1', 'peptide-2']})
table = features.export_features('amplitude', attributes=['name'])
print(table.shape)  # (2, 161): metadata + 20 channels × 8 unique Fourier bins

pairs = ComplexSet(method='z', transform_options={'radius': 1.05})
pairs.from_sequence_list(['ACDE'], ['ACDF'], length=8)
paired_table = pairs.export_features(('amplitude', 'phase'))
chain_a_table = pairs.export_features('amplitude', chain_ID='A')

result = transform_sequence('ACDE', method='fourier', length=8)
values, names = result.features(selector=lambda c: c['frequency'] >= 0)

The default BLOSUM100 embedding supplies 20 residue-property channels. Kidera, AAindex, and custom channel mappings work with every transform. Inputs are case-sensitive; unknown residues are rejected when not covered by the embedding. All chains in a collection share a fixed grid; no sequence truncation occurs. Per-sequence centering and none/length/sqrt_length/energy normalization operate on real residues before padding. Metadata cannot overwrite features.

component selects amplitude, phase, real, imaginary, or an ordered tuple; kinds selects coefficients and/or within-channel cross_coefficients. selector receives named coordinate arrays: frequency for Fourier/Z, log-frequency for Mellin, band/index for wavelets. Primary exports include all coefficients by default, with no periodic duplicate. Paired products are aligned coefficient descriptors, not a physical model of residue contacts.

For fitted pipelines use protspect.sklearn.SequenceTransformer, with the same method/options/components and coordinate selectors. Set paired=True for (A, B) samples. Grid and channel moments fit on training chains only; selection and output columns are frozen. See the unified API, transform definitions, and normalization formulas.

SequenceFFT, SpectralTransformer, and the sets' export_datatable methods remain compatibility adapters. Existing Fourier values, selectors, closure bins, and feature names are preserved. Use them with existing fitted models; recompute and refit before switching to the primary full-bin feature schema.

Migration from FFTpro

FFTpro.features, FFTpro.embeddings, and FFTpro.utility remain compatibility imports. New code should use protspect. The misspelled selector frequency_selelect_positive is retained as an alias.

Feature values have intentionally changed. The original implementation labeled unshifted FFT coefficients using a centered grid, assigning coefficients to incorrect frequencies. The corrected implementation aligns bins and labels. The BLOSUM100 P-to-A score was also corrected from +2 to -2, matching A-to-P. Regenerate existing feature tables and retrain downstream models; old and new features must not be mixed. Empty sequences, short FFT lengths, unequal pair counts, and unequal metadata lengths now raise errors instead of failing later or silently dropping records.

The historical notebook in FFTpro/examples/ is archived with all saved outputs cleared; the original CPP CSVs remain unchanged and excluded from distributions. A new reproducible CPP notebook and workflow uses the complete official MLCPP 2.0 website datasets and a pinned MISTIC source, a separate locked research environment, training-only model selection, exact train/test overlap checks, and recorded provenance. See executed results and official data preparation and provenance. Research dependencies are optional and separate from the runtime package. Use examples/basic_features.py for a minimal core-only example.

Development and release

python -m pytest -q
python -m ruff check .
python -m build
python -m twine check dist/*

For the controlled antimicrobial-peptide correction audit, see validation/README.md. Install .[dev,validation] to run its classifier tests and benchmark.

CI checks Python 3.10–3.15 on Linux, macOS, and Windows, installing both wheel and source archives into fresh environments outside the checkout. Exact minimum dependencies are checked on Python 3.10. See the compatibility policy for scope and evidence. See RELEASE_REVIEW.md for findings, validation evidence, and the remaining release decisions. The current alpha version is 0.1.0a2. The release record links the successful TestPyPI rehearsal and production download/install checks.

Citation and project contact

Cite the software version and commit used:

Biomolecular Systems Modeling and Design Group. Protspect: protein sequence spectral analysis. Version 0.1.0a2. https://github.com/biomodsquad/protein-sequence-spectral-analysis

Machine-readable metadata is in CITATION.cff. Cite dataset sources and external methods separately; the CPP example lists these in its provenance documentation.

The group maintains this package through GitHub Issues. See the changelog, support policy, and contribution guide.

Licensed under the BSD 3-Clause License.

Metadata

Release files for Protspect 0.1.0a2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for Protspect 0.1.0a2
File Size Uploaded Provenance Actions
protspect-0.1.0a2.tar.gz 1.6 MB With provenance Details for protspect-0.1.0a2.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for Protspect 0.1.0a2
File Interpreter ABI Platform Actions
protspect-0.1.0a2-py3-none-any.whl Python 3 none any Details for protspect-0.1.0a2-py3-none-any.whl

Total release size: 1.7 MB

Release files / protspect-0.1.0a2.tar.gz

Download URL protspect-0.1.0a2.tar.gz
Size 1.6 MB
Tags Source
SHA-256 checksum
How to use checksums
ab5ce90289560bdb91971dd4d9936e91e889df98844a2ba5c285aa74721d6c61
BLAKE2b-256 checksum
How to use checksums
a8dbd6b22e2f4bd08fa6a361e70705b3b419284b54785273793c5f69867ab742
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 27, 2026.

Transparency log

Release files / protspect-0.1.0a2-py3-none-any.whl

Download URL protspect-0.1.0a2-py3-none-any.whl
Size 161.8 kB
Tags Python 3
SHA-256 checksum
How to use checksums
5fe6e7a7c3431016d912c1a24c85b316f56220202c9b75b11ff24302295452cb
BLAKE2b-256 checksum
How to use checksums
94a3e9683bd677dab41100d5e5e47a35ae3a63ee59c760b112479ede15f229c7
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.16

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 27, 2026.

Transparency log

Security for Protspect 0.1.0a2

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for protspect-0.1.0a2.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 27, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page