protpy - Package for generating protein physicochemical, biochemical and structural descriptors using their constituent amino acids.

• 🧬 A demo of the software is available here
• 📝 A Medium article about protPy and its background is available here

Table of Contents

• Introduction
• Requirements
• Installation
• Usage
• Documentation
• Directories
• Tests
• Issues
• Contact
• References

Introduction

protpy is a Python software package for generating a variety of physicochemical, biochemical and structural descriptors for proteins. All of these descriptors are calculated using sequence-derived or physicochemical features of the amino acids that make up the proteins. These descriptors have been highly studied and used in a series of Bioinformatic applications including protein engineering, SAR (sequence-activity-relationships), predicting protein structure & function, subcellular localization, protein-protein interactions, drug-target interactions etc.

This software is aimed at any researcher or developer using protein sequence/structural data, and was mainly created to use in my own project pySAR which uses protein sequence data to identify Sequence Activity Relationships (SAR) using Machine Learning [1]. protpy is built and developed in Python 3.10.

The descriptors available in protpy include:

Composition Descriptors (24)

• amino_acid_composition (AAComp)
• dipeptide_composition (DPComp)
• gravy (GRAVY)
• aromaticity (Aromaticity)
• instability_index (II)
• isoelectric_point (pI)
• molecular_weight (MW)
• aliphatic_index (AI)
• extinction_coefficient (ExtCoeff)
• boman_index (BI) — alias of gravy
• charge_distribution (ChargeDist)
• hydrophobic_moment (μH)
• motif_composition (MotifComp)
• pseudo_amino_acid_composition (PAAComp)
• amphiphilic_pseudo_amino_acid_composition (APAAComp)
• natural_vector (NV)
• atomic_composition (ATC)
• cksaagp (CKSAAGP) — alias of cksaap
• cksaap (CKSAAP)
• asdc (ASDC)
• distance_pair (DP) — alias of cksaap
• tripeptide_composition (TPComp) — alias of kmer_composition
• kmer_composition (kmer)
• dde (DDE) — alias of dipeptide_composition

Complexity Descriptors (3)

• shannon_entropy (ShannonH)
• wootton_federhen_complexity (WF)
• lempel_ziv_complexity (LZ)

Autocorrelation Descriptors (4)

• geary_autocorrelation (GAuto) — alias of autocorrelation
• moreaubroto_autocorrelation (MBAuto) — alias of autocorrelation
• moran_autocorrelation (MAuto) — alias of autocorrelation
• autocorrelation (Auto)

Conjoint Triad Descriptors (1)

• conjoint_triad (CTriad)

CTD Descriptors (5)

• ctd (CTD)
• ctd_ (CTD) — alias of ctd
• ctd_composition (CTD-C) — alias of ctd
• ctd_transition (CTD-T) — alias of ctd
• ctd_distribution (CTD-D) — alias of ctd

Sequence Order Descriptors (2)

• quasi_sequence_order (QSO)
• sequence_order_coupling_number (SOCN)

Physicochemical Descriptors (5)

• charge_fractions (FCR)
• kappa (Kappa)
• sequence_charge_decoration (SCD)
• hydrophobicity_profile (HP)
• thermostability_indices (Thermo)

Spectral Descriptors (1)

• spectral_descriptors (Spectral)

Structural Descriptors (3)

• aggrescan (AGG)
• charge_hydropathy (CH)
• plaac (PLAAC)

Scales Descriptors (2)

• auto_cross_covariance (ACC)
• aa_scale_descriptors (Scale)

New in v1.6.0: a slimmer, corrected descriptor set

A descriptor value audit found that about half of the 80 registered descriptors were exact duplicates, subsets or fixed linear maps of the amino acid composition, and that several values were wrong. v1.6.0:

• fixes values — the instability-index table (now matches Biopython/ExPASy), QSO composition terms (were 100x too large), CTD distribution quartiles, FoldIndex residue assignment, kappa for single-sign sequences, APAAC denominator guard; pI and charge default to the Bjellqvist pKa set; conjoint triad is normalised as in Shen et al. 2007; descriptors are no longer rounded (decimals= rounds on request); a lag that is too long raises instead of silently becoming 30;
• merges duplicates into family functions — kmer_composition(alphabet=), cksaap(alphabet=), dipeptide_composition(normalize="dde"), autocorrelation(method=), ctd(parts=), sequence_order_coupling_number / quasi_sequence_order(distance_matrix="both"), hydrophobicity_profile(segments=), motif_composition (PROSITE motifs), aa_scale_descriptors(mode=), charge_hydropathy(window=), thermostability_indices() and spectral_descriptors(transform=); literature names such as tripeptide_composition, dde, cksaagp, distance_pair, geary_autocorrelation and ctd_composition stay as permanent aliases;
• deprecates 37 functions (they still work with a DeprecationWarning and are removed in 2.0; see the table in DESCRIPTORS.md);
• shrinks the default all_descriptors set from 9,957 to 1,402 columns (tripeptide composition is now opt-in), and makes it about 8x faster.

Category	Registered descriptors added in v1.5.0 / v1.6.0
Composition	cksaap, cksaagp, dde, asdc, distance_pair, natural_vector, atomic_composition
Autocorrelation / CTD	autocorrelation, ctd
Scales	aa_scale_descriptors, auto_cross_covariance
Physicochemical	charge_fractions, kappa, sequence_charge_decoration, hydrophobicity_profile, thermostability_indices
Complexity	wootton_federhen_complexity, lempel_ziv_complexity (and shannon_entropy, moved here)
Spectral	spectral_descriptors
Structural	aggrescan, charge_hydropathy, plaac

all_descriptors and encode_dataset take descriptor_set="classic" (default, 1,402 columns), "extended" or "all", plus include= / exclude= lists and decimals=:

df = protpy.all_descriptors(protein_seq, descriptor_set="extended")
df = protpy.all_descriptors(protein_seq, include=["tripeptide_composition"], decimals=3)

More detail of each descriptor is listed in the markdown file: DESCRIPTORS.md

Requirements

• Python >= 3.9
• aaindex >= 1.3.1
• numpy >= 2.0.0
• pandas >= 2.0.2
• tqdm >= 4.0.0
• biopython >= 1.87 (only required for testing)

Installation

Install the latest version of protpy using pip:

pip3 install protpy --upgrade

Install by cloning repository:

git clone https://github.com/amckenna41/protpy.git
python3 setup.py install

Usage

Import protpy after installation:

import protpy as protpy

Import protein sequence from fasta:

from Bio import SeqIO

with open("test_fasta.fasta") as pro:
    protein_seq = str(next(SeqIO.parse(pro,'fasta')).seq)

Composition Descriptors Usage Examples

Calculate Amino Acid Composition:

aac = protpy.amino_acid_composition(protein_seq)
# A      C      D      E      F ...
# 6.693  3.108  5.817  3.347  6.614 ...
# shape: 1 x 20

Calculate Dipeptide Composition:

dpc = protpy.dipeptide_composition(protein_seq)
# AA     AC     AD     AE     AF ...
# 0.797  0.159  0.478  0.399  0.239 ...
# shape: 1 x 400

#dipeptide deviation from expected mean (alias: protpy.dde)
dde = protpy.dipeptide_composition(protein_seq, normalize="dde")
# DDE_AA  DDE_AC  DDE_AD  DDE_AE  DDE_AF ...
# 1.989   -0.424  2.014   1.405   0.185 ...
# shape: 1 x 400

Calculate Tripeptide Composition:

#alias of kmer_composition(k=3); 8000 columns, not in any all_descriptors set
tpc = protpy.tripeptide_composition(protein_seq)
# AAA    AAC  AAD  AAE    AAF ...
# 0.080  0    0    0.160  0 ...
# shape: 1 x 8000

Calculate k-mer / Reduced Alphabet Composition:

#using default parameters: k=2
kmer = protpy.kmer_composition(protein_seq)
# AA     AC     AD     AE     AF ...
# 0.797  0.159  0.478  0.399  0.239 ...
# shape: 1 x 400

#hydrophobic / polar / charged groups (formerly hydrophobic_polar_charged_composition)
hpc = protpy.kmer_composition(protein_seq, k=1, alphabet=3)
# ReducedAlphabet_1  ReducedAlphabet_2  ReducedAlphabet_3
# 44.542             32.669             22.789
# shape: 1 x 3

#grouped dipeptide composition over the iFeature 5-group alphabet (GDPC)
gdpc = protpy.kmer_composition(protein_seq, k=2, alphabet="ifeature5")
# GDPC_aliphatic.aliphatic  GDPC_aliphatic.aromatic  GDPC_aliphatic.positive  GDPC_aliphatic.negative  GDPC_aliphatic.uncharged ...
# 12.839                    4.625                    3.429                    2.951                    12.121 ...
# shape: 1 x 25

#a custom grouping: a list of residue strings covering the 20 amino acids
custom = protpy.kmer_composition(protein_seq, k=1, alphabet=["AGV", "ILFP", "YMTS", "HNQW", "RK", "DE", "C"])
# RA_custom_AGV  RA_custom_ILFP  RA_custom_YMTS  RA_custom_HNQW  RA_custom_RK ...
# 20.239         25.259          21.434          12.908          7.888 ...
# shape: 1 x 7

Calculate GRAVY:

gravy = protpy.gravy(protein_seq)
# GRAVY
# -0.045
# shape: 1 x 1

Calculate Aromaticity:

aromaticity = protpy.aromaticity(protein_seq)
# Aromaticity
# 0.118
# shape: 1 x 1

Calculate Instability Index:

#score < 40 = stable protein; DIWV table as in Biopython / ExPASy
instability = protpy.instability_index(protein_seq)
# InstabilityIndex
# 32.419
# shape: 1 x 1

Calculate Isoelectric Point:

#using default parameters: pka_scale="bjellqvist" (as ExPASy / Biopython)
pi = protpy.isoelectric_point(protein_seq)
# IsoelectricPoint
# 5.562
# shape: 1 x 1

Calculate Molecular Weight:

mw = protpy.molecular_weight(protein_seq)
# MolecularWeight
# 139125.120
# shape: 1 x 1

Calculate Charge Distribution:

#using default parameters: ph=7.4, pka_scale="bjellqvist"
charge = protpy.charge_distribution(protein_seq)
# PositiveCharge  NegativeCharge  NetCharge
# 99.398          114.920         -15.522
# shape: 1 x 3

#charge per Da (formerly charge_density)
charge = protpy.charge_distribution(protein_seq, normalize="mass")
# PositiveCharge_perDa  NegativeCharge_perDa  NetCharge_perDa
# 0.000714              0.000826              -0.000112
# shape: 1 x 3

Calculate Motif Composition:

#using the 8 default motifs
motif = protpy.motif_composition(protein_seq)
# NxS/T_glycosylation  RGD_integrin  KDEL_retention  CxxC_zinc_finger  CAAX_prenylation ...
# 23                   0             0               2                 0 ...
# shape: 1 x 8

#built-in PROSITE motifs by name, with counts per 100 residues
motif = protpy.motif_composition(protein_seq, motifs=["N_glycosylation", "PTS1", "tryptic_site"], per_100=True)
# N_glycosylation  PTS1  tryptic_site  N_glycosylation_per100  PTS1_per100 ...
# 23               0     93            1.833                   0 ...
# shape: 1 x 6

Calculate Aliphatic Index:

ai = protpy.aliphatic_index(protein_seq)
# AliphaticIndex
# 82.725
# shape: 1 x 1

Calculate Extinction Coefficient:

ec = protpy.extinction_coefficient(protein_seq)
# ExtCoeff_Reduced  ExtCoeff_Oxidized
# 140960            143335
# shape: 1 x 2

Calculate Boman Index:

bi = protpy.boman_index(protein_seq)
# BomanIndex
# 1.198
# shape: 1 x 1

Calculate Hydrophobic Moment:

#using default parameters: window=11, angle=100 (alpha-helix)
hm = protpy.hydrophobic_moment(protein_seq)
# HydrophobicMoment_Mean  HydrophobicMoment_Max
# 0.272                   0.813
# shape: 1 x 2

Calculate Pseudo Amino Acid Composition:

#using default parameters: lamda=30, weight=0.05
paac = protpy.pseudo_amino_acid_composition(protein_seq)
# PAAC_1  PAAC_2  PAAC_3  PAAC_4  PAAC_5 ...
# 1.871   0.869   1.626   0.935   1.849 ...
# shape: 1 x 50

Calculate Amphiphilic Pseudo Amino Acid Composition:

#using default parameters: lamda=30, weight=0.5
apaac = protpy.amphiphilic_pseudo_amino_acid_composition(protein_seq)
# APAAC_1  APAAC_2  APAAC_3  APAAC_4  APAAC_5 ...
# 6.049    2.809    5.257    3.025    5.977 ...
# shape: 1 x 80

Calculate CKSAAP / CKSAAGP / ASDC / Distance Pair:

#using default parameters: gap=5
cksaap = protpy.cksaap(protein_seq)
# CKSAAP_AA_gap0  CKSAAP_AC_gap0  CKSAAP_AD_gap0  CKSAAP_AE_gap0  CKSAAP_AF_gap0 ...
# 0.797           0.159           0.478           0.399           0.239 ...
# shape: 1 x 2400

#grouped pairs (alias: protpy.cksaagp)
cksaagp = protpy.cksaap(protein_seq, alphabet="ifeature5")
# CKSAAGP_aliphatic.aliphatic_gap0  CKSAAGP_aliphatic.aromatic_gap0  CKSAAGP_aliphatic.positive_gap0  CKSAAGP_aliphatic.negative_gap0  CKSAAGP_aliphatic.uncharged_gap0 ...
# 12.839                            4.625                            3.429                            2.951                            12.121 ...
# shape: 1 x 150

asdc = protpy.asdc(protein_seq)
# ASDC_AA  ASDC_AC  ASDC_AD  ASDC_AE  ASDC_AF ...
# 0.443    0.218    0.329    0.252    0.329 ...
# shape: 1 x 400

dp = protpy.distance_pair(protein_seq)
# DP_IMV  DP_L   DP_F   DP_WY  DP_G ...
# 15.060  7.888  6.614  5.179  6.295 ...
# shape: 1 x 602

Calculate Natural Vector / Atomic Composition:

nv = protpy.natural_vector(protein_seq, normalize=True)
# NV_n_A  NV_n_C  NV_n_D  NV_n_E  NV_n_F ...
# 0.067   0.031   0.058   0.033   0.066 ...
# shape: 1 x 60

atc = protpy.atomic_composition(protein_seq)
# ATC_C   ATC_H   ATC_N  ATC_O   ATC_S
# 27.011  52.281  6.952  13.501  0.255
# shape: 1 x 5

Autocorrelation Descriptors Usage Examples

Calculate Autocorrelation:

#Moreau-Broto, Moran or Geary; the three names below are aliases
auto = protpy.autocorrelation(protein_seq, method="moran", lag=10)
# MAuto_CIDH920105_1  MAuto_CIDH920105_2  MAuto_CIDH920105_3  MAuto_CIDH920105_4  MAuto_CIDH920105_5 ...
# -0.058              -0.078              -0.042              -0.021              -0.054 ...
# shape: 1 x 80

#using default parameters: lag=30, the 8 default AAindex properties, normalize=True
mb = protpy.moreaubroto_autocorrelation(protein_seq)
# MBAuto_CIDH920105_1  MBAuto_CIDH920105_2  MBAuto_CIDH920105_3  MBAuto_CIDH920105_4  MBAuto_CIDH920105_5 ...
# -0.052               -0.072               -0.036               -0.015               -0.048 ...
# shape: 1 x 240

moran = protpy.moran_autocorrelation(protein_seq)
# MAuto_CIDH920105_1  MAuto_CIDH920105_2  MAuto_CIDH920105_3  MAuto_CIDH920105_4  MAuto_CIDH920105_5 ...
# -0.058              -0.078              -0.042              -0.021              -0.054 ...
# shape: 1 x 240

#Geary is the autocorrelation in the classic all_descriptors set
geary = protpy.geary_autocorrelation(protein_seq)
# GAuto_CIDH920105_1  GAuto_CIDH920105_2  GAuto_CIDH920105_3  GAuto_CIDH920105_4  GAuto_CIDH920105_5 ...
# 1.057               1.077               1.040               1.020               1.052 ...
# shape: 1 x 240

Conjoint Triad Descriptors Usage Examples

Calculate Conjoint Triad:

#normalised (f - min) / max as in Shen et al. 2007
ctriad = protpy.conjoint_triad(protein_seq)
# conj_triad_111  conj_triad_112  conj_triad_113  conj_triad_114  conj_triad_115 ...
# 0.409           0.773           0.500           0.136           0.273 ...
# shape: 1 x 343

#raw triad counts
ctriad = protpy.conjoint_triad(protein_seq, normalize=False)
# conj_triad_111  conj_triad_112  conj_triad_113  conj_triad_114  conj_triad_115 ...
# 9               17              11              3               6 ...
# shape: 1 x 343

CTD Descriptors Usage Examples

Calculate CTD:

#using default parameters: all 7 properties, parts="CTD" (ctd_ is an alias)
ctd = protpy.ctd(protein_seq)
# CTD_C_01_hydrophobicity  CTD_C_02_hydrophobicity  CTD_C_03_hydrophobicity  CTD_C_01_normalized_vdwv  CTD_C_02_normalized_vdwv ...
# 0.279                    0.386                    0.335                    0.420                     0.355 ...
# shape: 1 x 147

#transition only, for one property (same as ctd_transition)
ctd_t = protpy.ctd(protein_seq, property="charge", all_ctd=False, parts="T")
# CTD_T_12_charge  CTD_T_13_charge  CTD_T_23_charge
# 0.131            0.019            0.147
# shape: 1 x 3

Sequence Order Descriptors Usage Examples

Calculate Sequence Order Coupling Number:

#using default parameters: lag=30, distance_matrix="schneider-wrede"
socn = protpy.sequence_order_coupling_number(protein_seq)
# SOCN_SW1  SOCN_SW2  SOCN_SW3  SOCN_SW4  SOCN_SW5 ...
# 401.387   409.243   376.946   393.042   396.196 ...
# shape: 1 x 30

#both matrices, divided by the number of pairs (removes the length dependence)
socn = protpy.sequence_order_coupling_number(protein_seq, lag=10, distance_matrix="both", normalize=True)
# SOCN_SW1  SOCN_SW2  SOCN_SW3  SOCN_SW4  SOCN_SW5 ...
# 0.320     0.327     0.301     0.314     0.317 ...
# shape: 1 x 20

Calculate Quasi Sequence Order:

#using default parameters: lag=30, weight=0.1, distance_matrix="schneider-wrede"; the 50 values sum to 1
qso = protpy.quasi_sequence_order(protein_seq)
# QSO_SW1   QSO_SW2   QSO_SW3   QSO_SW4   QSO_SW5 ...
# 0.000057  0.000026  0.000049  0.000028  0.000056 ...
# shape: 1 x 50

#both matrices (as in all_descriptors)
qso = protpy.quasi_sequence_order(protein_seq, distance_matrix="both")
# QSO_SW1   QSO_SW2   QSO_SW3   QSO_SW4   QSO_SW5 ...
# 0.000057  0.000026  0.000049  0.000028  0.000056 ...
# shape: 1 x 100

Physicochemical, Complexity, Spectral and Structural Descriptors Usage Examples

Calculate Charge:

fcr = protpy.charge_fractions(protein_seq)
# Charge_FCR  Charge_FPlus  Charge_FMinus  Charge_NCPR  Charge_Sigma
# 0.171       0.079         0.092          -0.013       0.000953
# shape: 1 x 5

kappa = protpy.kappa(protein_seq)
# Kappa
# 0.153
# shape: 1 x 1

scd = protpy.sequence_charge_decoration(protein_seq)
# Charge_SCD
# 1.392
# shape: 1 x 1

Calculate Hydrophobicity Profile:

hp = protpy.hydrophobicity_profile(protein_seq)
# HP_kyte_doolittle_w9_mean  HP_kyte_doolittle_w9_max  HP_kyte_doolittle_w9_min  HP_kyte_doolittle_w9_std
# -0.053                     3.189                     -2.822                    0.876
# shape: 1 x 4

#hydropathy transmembrane segments (formerly transmembrane_features)
tm = protpy.hydrophobicity_profile(protein_seq, window=19, threshold=1.6, segments=True)
# HP_kyte_doolittle_w19_mean  HP_kyte_doolittle_w19_max  HP_kyte_doolittle_w19_min  HP_kyte_doolittle_w19_std  HP_kyte_doolittle_w19_frac_above ...
# -0.058                      2.811                      -1.911                     0.631                      0.015 ...
# shape: 1 x 8

#Vihinen flexibility profile (formerly flexibility)
flex = protpy.hydrophobicity_profile(protein_seq, scale="vihinen_flexibility", window=9, edge=0.25)
# HP_vihinen_flexibility_w9_mean  HP_vihinen_flexibility_w9_max  HP_vihinen_flexibility_w9_min  HP_vihinen_flexibility_w9_std
# 0.995                           1.060                          0.931                          0.019
# shape: 1 x 4

Calculate Thermostability Indices:

thermo = protpy.thermostability_indices(protein_seq)
# IVYWREL  CvP_Bias
# 0.330    -9.323
# shape: 1 x 2

Calculate Complexity:

h = protpy.shannon_entropy(protein_seq)
# ShannonEntropy
# 4.163
# shape: 1 x 1

wf = protpy.wootton_federhen_complexity(protein_seq)
# WF_K2_mean  WF_K2_min  WF_LowComplexity_frac
# 3.001       2.055      0.030
# shape: 1 x 3

lz = protpy.lempel_ziv_complexity(protein_seq)
# LZ_Normalized
# 0.860
# shape: 1 x 1

Calculate Spectral Descriptors:

fft = protpy.spectral_descriptors(protein_seq)
# FFT_eiip_bin1  FFT_eiip_bin2  FFT_eiip_bin3  FFT_eiip_bin4  FFT_eiip_bin5 ...
# 0.001361       0.001641       0.001838       0.001610       0.001683 ...
# shape: 1 x 10

dct = protpy.spectral_descriptors(protein_seq, transform="dct")
# DCT_kyte_doolittle_1  DCT_kyte_doolittle_2  DCT_kyte_doolittle_3  DCT_kyte_doolittle_4  DCT_kyte_doolittle_5 ...
# -1.812                1.051                 2.011                 1.613                 -3.117 ...
# shape: 1 x 10

Calculate Factor Scales:

#sequence mean of a factor scale
scales = protpy.aa_scale_descriptors(protein_seq, "atchley_factors")
# AF_1    AF_2   AF_3    AF_4   AF_5
# -0.095  0.080  -0.141  0.337  -0.122
# shape: 1 x 5

#auto/cross covariance of the 5 z-scales
acc = protpy.auto_cross_covariance(protein_seq)
# ACC_1_1_lag1  ACC_1_2_lag1  ACC_1_3_lag1  ACC_1_4_lag1  ACC_1_5_lag1 ...
# -0.494        -0.181        0.021         -0.101        -0.119 ...
# shape: 1 x 125

Calculate Structural:

agg = protpy.aggrescan(protein_seq)
# AGG_a3vSA  AGG_NnHS  AGG_AATr  AGG_THSAr  AGG_a4vAHS
# 0.072      4.701     0.230     0.155      0.409
# shape: 1 x 5

#Uversky charge-hydropathy score of the whole sequence
ch = protpy.charge_hydropathy(protein_seq)
# CH_HQ
# 0.215
# shape: 1 x 1

#FoldIndex profile
fi = protpy.charge_hydropathy(protein_seq, window=51)
# CH_w51_mean  CH_w51_max  CH_w51_min  CH_w51_std
# 0.186        0.582       -0.190      0.132
# shape: 1 x 4

plaac = protpy.plaac(protein_seq, fill_value=0.0)
# PLAAC_LLR  PLAAC_COREscore  PLAAC_PrD_length  PLAAC_has_PrD
# -1.578     0                0                 0
# shape: 1 x 4

### Scaling features for machine learning

protpy returns each descriptor on its natural scale (percentages, counts, indices, pH, Da, ...). If your classifier needs features on a common scale such as [0, 1], fit a scaler on your training data and reuse it for test and new sequences. This scales every column to the range seen in your data without leaking information from the test set:

from sklearn.preprocessing import MinMaxScaler

X_train = protpy.encode_dataset(train_seqs, output_file=None, verbose=False).set_index("sequence_id")
X_test = protpy.encode_dataset(test_seqs, output_file=None, verbose=False).set_index("sequence_id")

scaler = MinMaxScaler(clip=True).fit(X_train)   # fit on training data only
X_train_scaled = scaler.transform(X_train)      # every column in [0, 1]
X_test_scaled = scaler.transform(X_test)        # clip=True keeps unseen extremes in [0, 1]

StandardScaler is a common alternative, and tree-based models don't need scaling. When cross-validating, put the scaler in a scikit-learn Pipeline so it is re-fitted on each training fold. Impute the few NaN values that descriptor_set="extended" can produce (e.g. with SimpleImputer) before scaling.

Documentation

The documentation for protpy is hosted on ReadTheDocs and is available here.

Directories

• /tests - unit and integration tests for protpy package.
• /protpy - source code and all required external data files for package.
• /docs - protpy documentation.
• /examples - example notebook for protpy

Tests

Install the test requirements (these include Biopython, which the reference-value tests compare against), then run all tests from the main protpy folder:

pip install -r tests/requirements.txt
python3 -m unittest discover tests -v

python -m pytest runs the same suite, and the CI uses it. Set PROTPY_PERF=1 to include the timing test. See tests/README.md for what each test file covers.

Contact

If you have any questions or comments, please contact amckenna41@qub.ac.uk or raise an issue on the Issues tab.

References

[1]: Mckenna, A., & Dubey, S. (2022). Machine learning based predictive model for the analysis of sequence activity relationships using protein spectra and protein descriptors. Journal of Biomedical Informatics, 128(104016), 104016. https://doi.org/10.1016/j.jbi.2022.104016
[2]: Shuichi Kawashima, Minoru Kanehisa, AAindex: Amino Acid index database, Nucleic Acids Research, Volume 28, Issue 1, 1 January 2000, Page 374, https://doi.org/10.1093/nar/28.1.374
[3]: Dong, J., Yao, ZJ., Zhang, L. et al. PyBioMed: a python library for various molecular representations of chemicals, proteins and DNAs and their interactions. J Cheminform 10, 16 (2018). https://doi.org/10.1186/s13321-018-0270-2
[4]: Reczko, M. and Bohr, H. (1994) The DEF data base of sequence based protein fold class predictions. Nucleic Acids Res, 22, 3616-3619.
[5]: Hua, S. and Sun, Z. (2001) Support vector machine approach for protein subcellular localization prediction. Bioinformatics, 17, 721-728.
[6]: Broto P, Moreau G, Vandicke C: Molecular structures: perception, autocorrelation descriptor and SAR studies. Eur J Med Chem 1984, 19: 71–78.
[7]: Ong, S.A., Lin, H.H., Chen, Y.Z. et al. Efficacy of different protein descriptors in predicting protein functional families. BMC Bioinformatics 8, 300 (2007). https://doi.org/10.1186/1471-2105-8-300
[8]: Inna Dubchak, Ilya Muchink, Stephen R.Holbrook and Sung-Hou Kim. Prediction of protein folding class using global description of amino acid sequence. Proc.Natl. Acad.Sci.USA, 1995, 92, 8700-8704.
[9]: Juwen Shen, Jian Zhang, Xiaomin Luo, Weiliang Zhu, Kunqian Yu, Kaixian Chen, Yixue Li, Huanliang Jiang. Predicting proten-protein interactions based only on sequences inforamtion. PNAS. 2007 (104) 4337-4341.
[10]: Kuo-Chen Chou. Prediction of Protein Subcellar Locations by Incorporating Quasi-Sequence-Order Effect. Biochemical and Biophysical Research Communications 2000, 278, 477-483.
[11]: Kuo-Chen Chou. Prediction of Protein Cellular Attributes Using Pseudo-Amino Acid Composition. PROTEINS: Structure, Function, and Genetics, 2001, 43: 246-255.
[12]: Kuo-Chen Chou. Using amphiphilic pseudo amino acid composition to predict enzyme subfamily classes. Bioinformatics, 2005,21,10-19.

Support

Back to top