代谢组学数据分析

# 使用Scientific Skills获取代谢组学数据
# 技能：scientific-skills:metabolomics-workbench-database

# 检索心衰相关的代谢组学数据集
# 返回：数据集信息、样本数、代谢物数量

import metabolomics_workbench as mb

# 搜索心衰相关研究
studies = mb.search_studies(
  disease="heart failure",
  organism="Homo sapiens"
)

# 下载数据集
data = mb.download_study(study_id="ST000001")

# 查看数据维度
print(f"样本数: {data.n_samples}")
print(f"代谢物数: {data.n_metabolites}")

# 使用Scientific Skills查询HMDB数据库
# 技能：scientific-skills:hmdb-database

import hmdb_query as hmdb

# 注释代谢物信息
metabolites = ["Carnitine", "Acetylcarnitine", "Glutamate"]

for metab in metabolites:
  info = hmdb.get_metabolite_info(metab)
  print(f"代谢物: {info.name}")
  print(f" 化学分类: {info.classification}")
  print(f" 生物学功能: {info.biological_role}")
  print(f" 相关疾病: {info.diseases}")

# 使用scipy进行统计检验
# 技能：scientific-skills:scipy

from scipy import stats
import pandas as pd
import numpy as np

# 分离心衰组和对照组
hf_group = data[data.group == "heart_failure"]
control_group = data[data.group == "control"]

# 对每个代谢物进行Wilcoxon检验
results = []
for metab in data.metabolites:
  # Wilcoxon秩和检验
  statistic, p_value = stats.mannwhitneyu(
    hf_group[metab],
    control_group[metab],
    alternative='two-sided'
  )

  # 计算log2倍数变化
  log2fc = np.log2(
    hf_group[metab].mean() / control_group[metab].mean()
  )

  results.append({
    'metabolite': metab,
    'log2FC': log2fc,
    'p_value': p_value
  })

# FDR校正
from statsmodels.stats.multitest import multipletests
results_df = pd.DataFrame(results)
results_df['FDR'] = multipletests(
  results_df['p_value'],
  method='fdr_bh'
)[1]

# 筛选差异代谢物
significant = results_df[
  (abs(results_df['log2FC']) > 1) &
  (results_df['FDR'] < 0.05)
]

# 使用KEGG数据库进行通路分析
# 技能：scientific-skills:kegg-database

import kegg_query as kegg

# 获取差异代谢物列表
diff_metabs = significant['metabolite'].tolist()

# KEGG通路富集
pathways = kegg.enrich_metabolites(
  metabolites=diff_metabs,
  organism='hsa' # 人类
)

# 显示显著富集的通路
for pathway in pathways:
  if pathway.p_adjust < 0.05:
    print(f"通路: {pathway.name}")
    print(f" p值: {pathway.p_adjust:.4f}")
    print(f" 涉及代谢物: {pathway.metabolites}")