File size: 4,347 Bytes
593dcaa 8503206 2b126e8 8503206 5dd932f 8503206 5dd932f 8503206 5dd932f 8503206 36bfc91 8503206 36bfc91 8503206 36bfc91 8503206 36bfc91 5dd932f 36bfc91 8503206 36bfc91 8503206 36bfc91 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 |
from collections import defaultdict
from dotenv import load_dotenv
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain_openai import ChatOpenAI
from prompt import *
from utils import *
import os
import re
load_dotenv()
class Validation():
def __init__(self, llm):
if llm.startswith('gpt'):
self.llm = ChatOpenAI(temperature=0, model_name=llm)
elif llm.startswith('gemini'):
self.llm = ChatGoogleGenerativeAI(temperature=0, model=llm)
else:
self.llm = ChatOpenAI(temperature=0, model_name=llm, api_key=os.environ['PERPLEXITY_API_KEY'], base_url="https://api.perplexity.ai")
def validate(self, df, api):
df = df.fillna('')
df['Genes'] = df['Genes'].str.replace(' ', '').str.upper()
df['rsID'] = df['rsID'].str.replace(' ', '').str.lower()
# Check if there are multiple Genes
sym = [',', '/', '|', '-', '(', ')']
i = 0
while i < len(df):
gene = df.loc[i, 'Genes']
for s in sym:
if s in gene:
genes = gene.split(s)
df.loc[i + 0.1], df.loc[i + 0.9] = df.loc[i], df.loc[i]
df = df.sort_index().reset_index(drop=True)
df.loc[i + 1, 'Genes'], df.loc[i + 2, 'Genes'] = genes[0], s.join(genes[1:])
break
i += 1
# Check if there are multiple rsIDs
i = 0
while i < len(df):
rsid = df.loc[i, 'rsID']
if ',' in rsid:
rsids = rsid.split(',')
df.loc[i + 0.1], df.loc[i + 0.9] = df.loc[i], df.loc[i]
df = df.sort_index().reset_index(drop=True)
df.loc[i + 1, 'rsID'], df.loc[i + 2, 'rsID'] = rsids[0], s.join(rsids[1:])
i += 1
# Check if there is SNPs without 'rs'
for i in df.index:
safe = True
snp = df.loc[i, 'rsID']
snp = snp.replace('l', '1')
if re.fullmatch('rs(\d)+|', snp):
pass
elif re.fullmatch('ts(\d)+', snp):
snp = 'r' + snp[1:]
elif re.fullmatch('s(\d)+', snp):
snp = 'r' + snp
elif re.fullmatch('(\d)+', snp):
snp = 'rs' + snp
else:
safe = False
df = df.drop(i)
if safe:
df.loc[i, 'rsID'] = snp
df.reset_index(drop=True, inplace=True)
df_clean = df.copy()
# Validate genes and SNPs with APIs
if api:
dbsnp = {}
for i in df.index:
snp = df.loc[i, 'SNPs']
gene = df.loc[i, 'Genes']
if snp not in dbsnp:
try:
res = call(f'https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi?db=snp&retmode=json&id={snp[2:]}').json()['result'][snp[2:]]
if 'error' not in res:
dbsnp[snp].extend([r['name'] for r in res['genes']])
except Exception as e:
print("Error at API", e)
pass
dbsnp[snp] = list(set(dbsnp[snp]))
if gene not in dbsnp[snp]:
for other in permutate(gene):
if other in dbsnp[snp]:
df.loc[i, 'Genes'] = other
print(f'{gene} corrected to {other}')
break
else:
df = df.drop(i)
# Check with GWAS ground truth
for i in df.index:
gene = df.loc[i, 'Genes']
snp = df.loc[i, 'rsID']
perms = permutate(gene)
for perm in perms:
if perm in ground_truth and snp in ground_truth[perm]:
df.loc[i, 'Genes'] = perm
if gene != perm:
print(f'{gene} corrected to {perm} with {snp}')
else:
print(f'{gene} and {snp} safe')
break
else:
print(f'{gene} and {snp} not found')
df = df.drop(i)
df.reset_index(drop=True, inplace=True)
return df, df_clean |