3

[更新]

问题

我有两个数据库：

1：

1   Name: D-Tagatose 1,6-bisphosphate
2   Name: 1-Phosphatidyl-D-myo-inositol;: 1-Phosphatidyl-1D-myo-inositol;: 1-Phosphatidyl-    myo-inositol;: Phosphatidyl-1D-myo-inositol;: (3-Phosphatidyl)-1-D-inositol;: 1,2-Diacyl-sn-glycero-3-phosphoinositol;: Phosphatidylinositol
3   Name: Androstenedione;: Androst-4-ene-3,17-dione;: 4-Androstene-3,17-dione
4   Name: Spermine;: N,N'-Bis(3-aminopropyl)-1,4-butanediamine
5   Name: H+;: Hydron

2：

>  <NAME>    Benzaldehyde, 4-[(trimethylsilyl)oxy]-     >  <SYNONYMS>    Benzaldehyde, p-(trimethylsiloxy)-
>  <NAME>    Benzeneacetic acid, methyl ester           >  <SYNONYMS>    q qer
>  <NAME>    Cyclopropaneoctanoic acid, 2-[[2-[(2-ethylcyclopropyl)methyl]cyclopropyl]methyl]-, methyl ester    >  <SYNONYMS>    Methyl 8-[2-((2-[(2-ethylcyclopropyl)methyl]cyclopropyl)methyl)cyclopropyl]octanoate #
>  <NAME>    Mevalonic lactone, trimethylsilyl deriv.   >  <SYNONYMS>    Mevalonic lactone, trimethylsilyl
>  <NAME>    Benzeneacetic acid, phenylmethyl ester     >  <SYNONYMS>    Acetic acid, phenyl-, benzyl ester

期望的输出：

将数据库 2 中的名称或同义词与数据库 1 名称匹配。我们谈论的是化合物，因此化合物名称可能会发生细微的变化。这就是为什么我也使用链接的在线数据库进行匹配。

测试输入：

请参阅链接中的 excel 文件。数据

我试过什么？

  1. 仅匹配名称（您必须从 db 1 中的名称中减去“名称”字符串）

  2. 匹配部分名称 -> 显然不是化学名称匹配的最佳主意。

  3. 使用以下数据库进行匹配）

    车比

    NIST

    PubChem

小 R 输入：

输入 1

structure(c(">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", 
">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", 
">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", 
">  <NAME>", " Benzaldehyde, 4-[(trimethylsilyl)oxy]-", " Benzeneacetic acid, methyl ester", 
" Cyclopropaneoctanoic acid, 2-[[2-[(2-ethylcyclopropyl)methyl]cyclopropyl]methyl]-, methyl ester", 
" Mevalonic lactone, trimethylsilyl deriv.", " Benzeneacetic acid, phenylmethyl ester", 
" Butanoic acid, 3,3-dimethyl-, methyl ester", " Acetic acid, (4-(trifluoromethoxy)phenyl)methyl ester", 
" Phosphoramidothioic acid, O,S-dimethyl ester", " Octanoic acid, phenylmethyl ester", 
" Benzenepropanoic acid, methyl ester", " 2-Propenoic acid, 3-phenyl-, methyl ester", 
" Propanoic acid, 2-methyl-, phenylmethyl ester", " Acetic acid, (2,3-dichlorophenyl)methyl ester", 
" L-Methionine, methyl ester", " Butanoic acid, phenylmethyl ester", 
"<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", 
"<SYNONYMS>", ">  <SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", 
"<SYNONYMS>", "<SYNONYMS>", ">  <SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", 
" Benzaldehyde, p-(trimethylsiloxy)-", " Acetic acid, phenyl-, methyl ester", 
" Methyl 8-[2-((2-[(2-ethylcyclopropyl)methyl]cyclopropyl)methyl)cyclopropyl]octanoate #", 
" Mevalonic lactone, trimethylsilyl", " Acetic acid, phenyl-, benzyl ester", 
" Butyric acid, 3,3-dimethyl-, methyl ester", " NA", " Methamidophos", 
" Octanoic acid, benzyl ester", " Hydrocinnamic acid, methyl ester", 
" Cinnamic acid, methyl ester", " Isobutyric acid, benzyl ester", 
" NA", " Methyl 2-amino-4-(methylsulfanyl)butanoate #", " Butyric acid, benzyl ester"
), .Dim = c(15L, 4L), .Dimnames = list(c("1", "2", "3", "4", 
"5", "6", "7", "8", "9", "10", "11", "12", "13", "14", "15"), 
    c("NAME", NA, "NA.1", "NA.2")))

输入 2

structure(c("Name: 1-Phosphatidyl-D-myo-inositol;: 1-Phosphatidyl-1D-myo-inositol;: 1-Phosphatidyl-myo-inositol;: Phosphatidyl-1D-myo-inositol;: (3-Phosphatidyl)-1-D-inositol;: 1,2-Diacyl-sn-glycero-3-phosphoinositol;: Phosphatidylinositol", 
"Name: Androstenedione;: Androst-4-ene-3,17-dione;: 4-Androstene-3,17-dione", 
"Name: Spermine;: N,N'-Bis(3-aminopropyl)-1,4-butanediamine", 
"Name: H+;: Hydron", "Name: 3-Iodo-L-tyrosine", "Name: 3-Methoxytyramine", 
"Name: 3-Methoxy-4-hydroxyphenylacetaldehyde;: (4-Hydroxy-3-methoxyphenyl)acetaldehyde;: Homovanillin", 
"Name: L-Noradrenaline;: Noradrenaline;: Norepinephrine;: Arterenol;: 4-[(1R)-2-Amino-1-hydroxyethyl]-1,2-benzenediol", 
"Name: 3,4-Dihydroxymandelaldehyde;: 3,4-Dihydroxyphenylglycolaldehyde", 
"Name: L-Metanephrine", "Name: L-Adrenaline;: (R)-(-)-Adrenaline;: (R)-(-)-Epinephrine;: (R)-(-)-Epirenamine;: (R)-(-)-Adnephrine;: 4-[(1R)-1-Hydroxy-2-(methylamino)ethyl]-1,2-benzenediol", 
"Name: 3-Methoxy-4-hydroxyphenylglycolaldehyde", "Name: L-Normetanephrine", 
"Name: L-Dopachrome;: 2-L-Carboxy-2,3-dihydroindole-5,6-quinone", 
"Name: 5,6-Dihydroxyindole;: DHI"), .Dim = c(15L, 1L))
4

2 回答 2

3

我认为问题没有简单的解决方案。使用微笑或 inchi/inchikeys 可能会很痛苦，匹配常见或 IUPAC 名称并不容易。

您可以检查 Pubchem PUG 方法以按名称检索化合物： https ://pubchem.ncbi.nlm.nih.gov/pug_rest/PUG_REST_Tutorial.html#_Toc338920590

Python 中的简单解决方案（在 R 中应该很容易实现）如下所示：

我将您的数据导出到纯文本文件：

输入1.csv

Benzaldehyde, 4-[(trimethylsilyl)oxy]-; Benzaldehyde, p-(trimethylsiloxy)-
Benzeneacetic acid, methyl ester; Acetic acid, phenyl-, methyl ester
Cyclopropaneoctanoic acid, 2-[[2-[(2-ethylcyclopropyl)methyl]cyclopropyl]methyl]-, methyl ester; Methyl 8-[2-((2-[(2-ethylcyclopropyl)methyl]cyclopropyl)methyl)cyclopropyl]octanoate #
Mevalonic lactone, trimethylsilyl deriv.; Mevalonic lactone, trimethylsilyl
Benzeneacetic acid, phenylmethyl ester; Acetic acid, phenyl-, benzyl ester
Butanoic acid, 3,3-dimethyl-, methyl ester; Butyric acid, 3,3-dimethyl-, methyl ester
Acetic acid, (4-(trifluoromethoxy)phenyl)methyl ester; NA
Phosphoramidothioic acid, O,S-dimethyl ester; Methamidophos
Octanoic acid, phenylmethyl ester; Octanoic acid, benzyl ester
Benzenepropanoic acid, methyl ester; Hydrocinnamic acid, methyl ester
2-Propenoic acid, 3-phenyl-, methyl ester; Cinnamic acid, methyl ester
Propanoic acid, 2-methyl-, phenylmethyl ester; Isobutyric acid, benzyl ester
Acetic acid, (2,3-dichlorophenyl)methyl ester; NA
L-Methionine, methyl ester; Methyl 2-amino-4-(methylsulfanyl)butanoate #
Butanoic acid, phenylmethyl ester; Butyric acid, benzyl ester

输入2.csv：

Name: 1-Phosphatidyl-D-myo-inositol;: 1-Phosphatidyl-1D-myo-inositol;: 1-Phosphatidyl-myo-inositol;: Phosphatidyl-1D-myo-inositol;: (3-Phosphatidyl)-1-D-inositol;: 1,2-Diacyl-sn-glycero-3-phosphoinositol;: Phosphatidylinositol
Name: Androstenedione;: Androst-4-ene-3,17-dione;: 4-Androstene-3,17-dione
Name: Spermine;: N,N'-Bis(3-aminopropyl)-1,4-butanediamine
Name: H+;: Hydron
Name: 3-Iodo-L-tyrosine
Name: 3-Methoxytyramine
Name: 3-Methoxy-4-hydroxyphenylacetaldehyde;: (4-Hydroxy-3-methoxyphenyl)acetaldehyde;: Homovanillin
Name: L-Noradrenaline;: Noradrenaline;: Norepinephrine;: Arterenol;: 4-[(1R)-2-Amino-1-hydroxyethyl]-1,2-benzenediol
Name: 3,4-Dihydroxymandelaldehyde;: 3,4-Dihydroxyphenylglycolaldehyde
Name: L-Metanephrine
Name: L-Adrenaline;: (R)-(-)-Adrenaline;: (R)-(-)-Epinephrine;: (R)-(-)-Epirenamine;: (R)-(-)-Adnephrine;: 4-[(1R)-1-Hydroxy-2-(methylamino)ethyl]-1,2-benzenediol
Name: 3-Methoxy-4-hydroxyphenylglycolaldehyde
Name: L-Normetanephrine
Name: L-Dopachrome;: 2-L-Carboxy-2,3-dihydroindole-5,6-quinone
Name: 5,6-Dihydroxyindole;: DHI

Python代码：

import requests

def name_to_cids(name):
    '''Retrive set of pubchem cids for given name'''
    url = 'https://pubchem.ncbi.nlm.nih.gov/rest/pug/compound/name/{0}/cids/JSON'.format(name)
    r = requests.get(url)
    return set(r.json()["IdentifierList"]["CID"])  if r.status_code == 200 else set()


def names_to_cids(names):
    '''Take list of names and return set of pubchem cids'''
    cids = set()
    for name in names:
        cids = cids.union(name_to_cids(name))
    return cids

def find_matching(from_key, from_dict, to_dict):
    matching = []
    from_cids = from_dict[from_key]['cids']
    for k, v in to_dict.items():
        if len(from_cids.intersection(v['cids'])) != 0:
            matching.append(k)
    return matching

#Read input files
input_1 = [line.replace('#','').strip().split(';') for line in open('input1.csv').readlines()]
input_2 = [[n for n in line.replace('Name: ', '').strip().split(';:') if n != 'NA'] for line in open('input2.csv').readlines()]

input_1_dict = {" ".join(names): {'names': names, 'cids': names_to_cids(names)} for names in input_1}
input_2_dict = {" ".join(names): {'names': names, 'cids': names_to_cids(names)} for names in input_2}

print(find_matching(from_key=input_1_dict.keys()[0], from_dict=input_1_dict, to_dict=input_2_dict))

Pubchem 不能很好地处理修改（例如，您可以找到苯乙酸的 cid，但不能找到苯乙酸甲酯），因此根据您的需要，您可以考虑删除查询字符串的某些部分（即搜索苯乙酸而不是苯乙酸，甲酯），但我理解它远非好。

更新

你也可以尝试一些更复杂的东西。

  • 一开始我会尝试确定您拥有的记录类别（通用名称，IUPAC 名称）。
  • 对于每个名称，您可以尝试使用正确的服务（Pubchem、ChEBI）或 OPSIN 等软件（https://bitbucket.org/dan2097/opsin）检索化学结构，具体取决于名称/标识符类别
  • 使用您选择的软件计算分子之间的成对 tanimoto 相似性。
  • 基于 tanimoto 系数识别您正在寻找的条目。
于 2013-09-06T13:55:41.387 回答
1

如果你真的想要一个 R 解决方案，试试下面的方法。我真的认为你需要整理你的输入，尤其是第二组的第四个元素。我所放的内容仅适用于每种化学品的名字。我会把它留给你处理同义词。

并非您的所有化学名称似乎都在 chemspider 数据库中有条目，您可能与其他人有更多的运气。捕捉那些没有名字的条目是函数的一个重要部分，没有它，一切都会中断。

您需要向 chemspider 注册以获取 api 的令牌。这样做是免费的。

您给出的示例化学名称在两个数据集之间似乎不匹配，因此下面的 df3 将不包含任何匹配项。我希望这有帮助。

library(RCurl)
library(XML)

token <- "your token here" # from chemspider profile
#url <- "http://www.chemspider.com/Search.asmx/AsyncSimpleSearch?query="
url <- "www.chemspider.com/Search.asmx/SimpleSearch?query="

chemCrawl <- function(chemname){ # Query chemspider with chemical names, return ids. 
  # df1[13] in particular seems to throw an error. Don't know why. 
  chem.id <-tryCatch(xmlValue(xmlRoot(xmlTreeParse(
    getURL(paste(url, "\"", curlEscape(chemname), "\"" ,"&token=" ,
                 token, sep = ""))
  ))), error=function(err) { 
    "oops"} )
  return(chem.id)
}

df1 <- as.data.frame(structure(c(">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", 
            ">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", 
            ">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", ">  <NAME>", 
            ">  <NAME>", " Benzaldehyde, 4-[(trimethylsilyl)oxy]-", " Benzeneacetic acid, methyl ester", 
            " Cyclopropaneoctanoic acid, 2-[[2-[(2-ethylcyclopropyl)methyl]cyclopropyl]methyl]-, methyl ester", 
            " Mevalonic lactone, trimethylsilyl deriv.", " Benzeneacetic acid, phenylmethyl ester", 
            " Butanoic acid, 3,3-dimethyl-, methyl ester", " Acetic acid, (4-(trifluoromethoxy)phenyl)methyl ester", 
            " Phosphoramidothioic acid, O,S-dimethyl ester", " Octanoic acid, phenylmethyl ester", 
            " Benzenepropanoic acid, methyl ester", " 2-Propenoic acid, 3-phenyl-, methyl ester", 
            " Propanoic acid, 2-methyl-, phenylmethyl ester", " Acetic acid, (2,3-dichlorophenyl)methyl ester", 
            " L-Methionine, methyl ester", " Butanoic acid, phenylmethyl ester", 
            "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", 
            "<SYNONYMS>", ">  <SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", 
            "<SYNONYMS>", "<SYNONYMS>", ">  <SYNONYMS>", "<SYNONYMS>", "<SYNONYMS>", 
            " Benzaldehyde, p-(trimethylsiloxy)-", " Acetic acid, phenyl-, methyl ester", 
            " Methyl 8-[2-((2-[(2-ethylcyclopropyl)methyl]cyclopropyl)methyl)cyclopropyl]octanoate #", 
            " Mevalonic lactone, trimethylsilyl", " Acetic acid, phenyl-, benzyl ester", 
            " Butyric acid, 3,3-dimethyl-, methyl ester", " NA", " Methamidophos", 
            " Octanoic acid, benzyl ester", " Hydrocinnamic acid, methyl ester", 
            " Cinnamic acid, methyl ester", " Isobutyric acid, benzyl ester", 
            " NA", " Methyl 2-amino-4-(methylsulfanyl)butanoate #", " Butyric acid, benzyl ester"
), .Dim = c(15L, 4L), .Dimnames = list(c("1", "2", "3", "4", 
                                         "5", "6", "7", "8", "9", "10", "11", "12", "13", "14", "15"), 
                                       c("NAME", NA, "NA.1", "NA.2"))))

names(df1) <-c("class1", "name", "class2", "synonym") 
df1$name <- as.character(df1$name)
df1[1,2] # there are leading spaces
df1$name <- sub(" ", "", df1$name) # lose the leading space
#details of chemspider search api: http://www.chemspider.com/Search.asmx

df1$chem.id <- lapply(df1$name, chemCrawl)
head(df1)

name2 <- structure(c("Name: 1-Phosphatidyl-D-myo-inositol;: 1-Phosphatidyl-1D-myo-inositol;: 1-Phosphatidyl-myo-inositol;: Phosphatidyl-1D-myo-inositol;: (3-Phosphatidyl)-1-D-inositol;: 1,2-Diacyl-sn-glycero-3-phosphoinositol;: Phosphatidylinositol", 
                   "Name: Androstenedione;: Androst-4-ene-3,17-dione;: 4-Androstene-3,17-dione", 
                   "Name: Spermine;: N,N'-Bis(3-aminopropyl)-1,4-butanediamine", 
                   "Name: H+;: Hydron", "Name: 3-Iodo-L-tyrosine", "Name: 3-Methoxytyramine", 
                   "Name: 3-Methoxy-4-hydroxyphenylacetaldehyde;: (4-Hydroxy-3-methoxyphenyl)acetaldehyde;: Homovanillin", 
                   "Name: L-Noradrenaline;: Noradrenaline;: Norepinephrine;: Arterenol;: 4-[(1R)-2-Amino-1-hydroxyethyl]-1,2-benzenediol", 
                   "Name: 3,4-Dihydroxymandelaldehyde;: 3,4-Dihydroxyphenylglycolaldehyde", 
                   "Name: L-Metanephrine", "Name: L-Adrenaline;: (R)-(-)-Adrenaline;: (R)-(-)-Epinephrine;: (R)-(-)-Epirenamine;: (R)-(-)-Adnephrine;: 4-[(1R)-1-Hydroxy-2-(methylamino)ethyl]-1,2-benzenediol", 
                   "Name: 3-Methoxy-4-hydroxyphenylglycolaldehyde", "Name: L-Normetanephrine", 
                   "Name: L-Dopachrome;: 2-L-Carboxy-2,3-dihydroindole-5,6-quinone", 
                   "Name: 5,6-Dihydroxyindole;: DHI"), .Dim = c(15L, 1L))

name2 <- sub("Name: ", "", name2)
name2 <- sub(";.+$", "", name2)
chem.id <- rep(NA, 15)
df2 <- as.data.frame(cbind(name2, chem.id))
names(df2)[1] <- "name2"
df2$chem.id <- lapply(df2$name2, chemCrawl)
head(df2)
df1$chem.id <- as.character(df1$chem.id)
df2$chem.id <- as.character(df2$chem.id)
df3 <- merge(df1, df2, by = "chem.id", all = TRUE)
df3
于 2013-09-07T16:24:08.510 回答