GitLocker: The Coding Marketplace

Description:

pandaspgs 0.1.0

pandasPGS: a Python package for easy retrieval of PGS Catalog data
Installation
pip install pandaspgs
Documentation
See pandasPGS Documentation
Licensing information
Source code
MIT License
Data from PGS Catalog
The PGS Catalog and all its contents are available under the general terms of use for EMBL-EBI services
Example 1. Investigating trends in diabetes-related polygenic risk scores
Step 1
from pandaspgs import *
from plotnine import*

Step 2
traits = get_traits(term='diabetes')
traits
# Trait is running in fat mode. It has 6 DataFrames with hierarchical dependencies.
# traits: 6 rows
# |
# -associated_pgs_ids: 186 rows
# |
# -child_associated_pgs_ids:265 rows
# |
# -trait_categories: 13 rows
# |
# -trait_mapped_terms: 57 rows
# |
# -trait_synonyms: 66 rows

Step 3
traits.traits
# id label description url
# 0 EFO_0000400 diabetes mellitus A metabolic disorder characterized by abnormal... http://www.ebi.ac.uk/efo/EFO_0000400
# 1 EFO_0006842 diabetes mellitus biomarker http://www.ebi.ac.uk/efo/EFO_0006842
# 2 EFO_0003770 diabetic retinopathy A chronic, pathological complication associate... http://www.ebi.ac.uk/efo/EFO_0003770
# 3 EFO_0004593 gestational diabetes Carbohydrate intolerance first diagnosed durin... http://www.ebi.ac.uk/efo/EFO_0004593
# 4 MONDO_0005147 type 1 diabetes mellitus A chronic condition characterized by minimal o... http://purl.obolibrary.org/obo/MONDO_0005147
# 5 MONDO_0005148 type 2 diabetes mellitus A type of diabetes mellitus that is characteri... http://purl.obolibrary.org/obo/MONDO_0005148

Step 4
score1 = get_scores(trait_id='EFO_0000400')
score2 = get_scores(trait_id='EFO_0006842')
score3 = get_scores(trait_id='EFO_0003770')
score4 = get_scores(trait_id='EFO_0004593')
score5 = get_scores(trait_id='MONDO_0005147')
score6 = get_scores(trait_id='MONDO_0005148')
diabetes_score = score1+score2+score3+score4+score5+score6
diabetes_score
# Score is running in fat mode. It has 7 DataFrames with hierarchical dependencies.
# scores:186 rows
# |
# -samples_variants: 253 rows
# |
# -samples_variants_cohorts: 386 rows
# |
# -samples_training: 107 rows
# |
# -samples_training_cohorts: 97 rows
# |
# -trait_efo: 195 rows
# |
# -ancestry_distribution: 447 rows

Step 5
pic=ggplot(diabetes_score.trait_efo)+geom_bar(aes(x='label'))+coord_flip()
pic.save(filename='Additional file 1.png',dpi=300)

Example 2: Investigating polygenic risk scores for gestational diabetes
Step 1
from pandaspgs import *

Step 2
traits = get_traits(term='gestational diabetes')
traits.traits
# id label description url
# 0 EFO_0004593 gestational diabetes Carbohydrate intolerance first diagnosed durin... http://www.ebi.ac.uk/efo/EFO_0004593
traits.traits['id'][0]
# 'EFO_0004593'
traits.traits['description'][0]
# 'Carbohydrate intolerance first diagnosed during pregnancy. [NCIT: P378]'

Step 3
gd_pgs = get_scores(trait_id='EFO_0004593')
gd_pgs.scores
# id name ftp_scoring_file matches_publication trait_reported trait_additional method_name method_params variants_number variants_interactions variants_genomebuild weight_type date_release license publication.id publication.title publication.doi publication.PMID publication.journal publication.firstauthor publication.date_publication ftp_harmonized_scoring_files.GRCh38.positions ftp_harmonized_scoring_files.GRCh37.positions
# 0 PGS002256 GRS4_GDM https://ftp.ebi.ac.uk/pub/databases/spot/pgs/s... True Gestational diabetes mellitus in early pregnancy None Genome-wide significant variants p < 0.05 4 0 NR log(OR) 2022-02-16 PGS obtained from the Catalog should be cited ... PGP000282 An early prediction model for gestational diab... 10.1186/s13098-022-00788-y 35073990 Diabetol Metab Syndr Wu Q 2022-01-24 https://ftp.ebi.ac.uk/pub/databases/spot/pgs/s... https://ftp.ebi.ac.uk/pub/databases/spot/pgs/s...
gd_pgs.scores['id'][0]
# PGS002256
gd_pgs.scores['name'][0]
# 'GRS4_GDM'
gd_pgs.scores['matches_publication'][0]
# True
gd_pgs.scores['trait_reported'][0]
# 'Gestational diabetes mellitus in early pregnancy'
gd_pgs.scores['variants_number'][0]
# 4

Step 4
gd_pgs.scores['publication.id'][0]
# PGP000282
gd_pgs.scores['publication.PMID'][0]
# 35073990
gd_pgs.scores['publication.date_publication'][0]
# '2022-01-24'
gd_pgs.scores['publication.journal'][0]
# 'Diabetol Metab Syndr'
gd_pgs.scores['publication.title'][0]
# 'An early prediction model for gestational diabetes mellitus based on genetic variants and clinical characteristics in China.'
open_in_pubmed(gd_pgs.scores['publication.PMID'][0])

Step 5
gd_pgs.samples_variants
# sample_number sample_cases sample_controls sample_percent_male sample_age phenotyping_free followup_time ancestry_broad ancestry_free ancestry_country ancestry_additional source_GWAS_catalog source_PMID source_DOI cohorts_additional id score_id followup_time.estimate_type followup_time.estimate followup_time.interval.type followup_time.interval.lower followup_time.interval.upper followup_time.variability_type followup_time.variability followup_time.unit
# 0 671 332 339 0.0 None None None East Asian Chinese China None None None None None 0 PGS002256 NaN NaN NaN NaN NaN NaN NaN NaN
gd_pgs.samples_variants['sample_number'][0]
# 671
gd_pgs.samples_variants['ancestry_broad'][0]
# 'East Asian'

Step 6
gd_file = read_scoring_file('PGS002256')
gd_file
# rsID effect_allele other_allele effect_weight hm_source hm_rsID hm_chr hm_pos hm_inferOtherAllele
# 0 rs10830963 G C 1.327 ENSEMBL rs10830963 11 92708710 NaN
# 1 rs1436953 T C 1.292 ENSEMBL rs1436953 15 62414014 NaN
# 2 rs7172432 G A 1.283 ENSEMBL rs7172432 15 62396389 NaN
# 3 rs16955379 C T 1.220 ENSEMBL rs16955379 16 81489373 NaN

Step 7
snp1=gd_file[['rsID','effect_allele','other_allele','effect_weight']].loc[0]
snp1
# rsID rs10830963
# effect_allele G
# other_allele C
# effect_weight 1.327
# Name: 0, dtype: object
from pandaspgs.file_operation import genotype_weighted_score
genotype_weighted_score(snp1)
# rs10830963_genotype rs10830963_weighted_score
# 0 G/G 2.654
# 1 G/C 1.327
# 2 C/C 0.000

Step 8
df_list = []
for x in range(len(gd_file)):
snp_x = gd_file[['rsID','effect_allele','other_allele','effect_weight']].loc[x]
df_x = genotype_weighted_score(snp_x)
df_x['key'] = 1
df_list.append(df_x)
from functools import reduce
combination_df = reduce(lambda x, y: x.merge(y,on='key'), df_list)
del combination_df['key']
combination_df
# rs10830963_genotype rs10830963_weighted_score rs1436953_genotype rs1436953_weighted_score rs7172432_genotype rs7172432_weighted_score rs16955379_genotype rs16955379_weighted_score
# 0 G/G 2.654 T/T 2.584 G/G 2.566 C/C 2.44
# 1 G/G 2.654 T/T 2.584 G/G 2.566 C/T 1.22
# 2 G/G 2.654 T/T 2.584 G/G 2.566 T/T 0.00
# 3 G/G 2.654 T/T 2.584 G/A 1.283 C/C 2.44
# 4 G/G 2.654 T/T 2.584 G/A 1.283 C/T 1.22
# .. ... ... ... ... ... ... ... ...
# 76 C/C 0.000 C/C 0.000 G/A 1.283 C/T 1.22
# 77 C/C 0.000 C/C 0.000 G/A 1.283 T/T 0.00
# 78 C/C 0.000 C/C 0.000 A/A 0.000 C/C 2.44
# 79 C/C 0.000 C/C 0.000 A/A 0.000 C/T 1.22
# 80 C/C 0.000 C/C 0.000 A/A 0.000 T/T 0.00
# [81 rows x 8 columns]

Step 9
combination_df['genotypes']=combination_df['rs10830963_genotype']+"-"+combination_df['rs1436953_genotype']+"-"+combination_df['rs7172432_genotype']+"-"+combination_df['rs16955379_genotype']
combination_df['scores']= combination_df['rs10830963_weighted_score']+combination_df['rs1436953_weighted_score']+combination_df['rs7172432_weighted_score']+combination_df['rs16955379_weighted_score']
combination_df[['genotypes','scores']].sort_values(by='scores', ascending=False)
# genotypes scores
# 0 G/G-T/T-G/G-C/C 10.244
# 1 G/G-T/T-G/G-C/T 9.024
# 3 G/G-T/T-G/A-C/C 8.961
# 9 G/G-T/C-G/G-C/C 8.952
# 27 G/C-T/T-G/G-C/C 8.917
# .. ... ...
# 53 G/C-C/C-A/A-T/T 1.327
# 71 C/C-T/C-A/A-T/T 1.292
# 77 C/C-C/C-G/A-T/T 1.283
# 79 C/C-C/C-A/A-C/T 1.220
# 80 C/C-C/C-A/A-T/T 0.000
#
# [81 rows x 2 columns]