【1】蛋白鉴定软件之X!Tandem

1. 简介

X!Tandem是GPM：The Global Proteome Machine（主要基于Web的开源用户界面，用于分析和显示蛋白质鉴定数据。关于GPM的更多内容可参考https://www.thegpm.org/GPM/faq.html）中的一个成员，此外还有X! P3 、X! Hunter等工具。顾名思义，X！表示X! series搜索引擎，Tandem表示串联质谱，所以X!Tandem用于串联质谱数据分析。

软件优点：

免费开源，win/linux/mac版本
集群系统可采用Parallel Tandem 或 X!!Tandem并行处理
运行速度还可以，c++写的
使用简单：命令行调用XML文件，输出格式（也是XML）适用于GPM其他工具
统计可靠：除了谱图匹配肽段序列，它还将数据中的所有肽段重新归并到已知蛋白序列，并且给出统计证据（组装和匹配是非随机的）。
不需要PeptideProphet 和 ProteinProphet等额外蛋白归并和统计分析软件。
蛋白期望值计算公式：

官网只提优点，没说缺点。根据试用的情况，简单提几点：

结果不友好：还需要后处理
软件更新慢：最新版ALANINE (2017.02.01)

2.下载安装

下载地址：
ftp://ftp.thegpm.org/projects/tandem/source
这里只试用Linux版本，下载解压

tar -xzf tandem-linux-17-02-01-4.zip

解压后三个子文件夹：

src包含软件源码
bin包含软件二进制文件和示例文件
fasta包含示例蛋白序列文件

3. 软件试用

软件运行很简单：

/path/bin/tandem.exe  input.xml

input.xml配置文件中调用了default_input.xml和taxonomy.xml文件，然后就是物种名，输入的谱图文件（需要mgf格式）和输出文件：

<?xml version="1.0"?>
<bioml>
    <note>
    Each one of the parameters for x! tandem is entered as a labeled note node.
    Any of the entries in the default_input.xml file can be over-ridden by
    adding a corresponding entry to this file. This file represents a minimum
    input file, with only entries for the default settings, the output file
    and the input spectra file name.
    See the taxonomy.xml file for a description of how FASTA sequence list
    files are linked to a taxon name.
    </note>
    <note type="input" label="list path, default parameters">default_input.xml</note>
    <note type="input" label="list path, taxonomy information">taxonomy.xml</note>
    <note type="input" label="protein, taxon">yeast</note>
    <note type="input" label="spectrum, path">test_spectra.mgf</note>
    <note type="input" label="output, path">output.xml</note>
</bioml>

示例中taxonomy.xml文件有多个物种的序列。而实际项目中我们往往针对一个物种，因此下载fasta文件配置一个即可：

<?xml version="1.0"?><bioml label="x! taxon-to-file matching list">
    <taxon label="yeast">
        <file format="peptide" URL="../fasta/scd.fasta.pro" />
        <file format="peptide" URL="../fasta/scd_1.fasta.pro" />
        <file format="peptide" URL="../fasta/crap.fasta.pro" />
    </taxon></bioml>

串联质谱的参数设置主要在default_input.xml中，可配置的参数很多，具体解释可参考https://www.thegpm.org/TANDEM/api/index.html。大部分使用默认即可，我们需要设置往往只有少数几个，比如一二级误差及其单位，可变修饰和固定修饰（注意修饰格式的写法与其他软件不同，更多的修饰可查看https://www.unimod.org/login.php?message=expired），漏切数等等。

<?xml version="1.0"?><?xml-stylesheet type="text/xsl" href="tandem-input-style.xsl"?><bioml><note>list path parameters</note>
    <note type="input" label="list path, default parameters">default_input.xml</note>
        <note>This value is ignored when it is present in the default parameter
        list path.</note>
    <note type="input" label="list path, taxonomy information">taxonomy.xml</note><note>spectrum parameters</note>
    <note type="input" label="spectrum, fragment monoisotopic mass error">0.4</note>
    <note type="input" label="spectrum, parent monoisotopic mass error plus">100</note>
    <note type="input" label="spectrum, parent monoisotopic mass error minus">100</note>
    <note type="input" label="spectrum, parent monoisotopic mass isotope error">yes</note>
    <note type="input" label="spectrum, fragment monoisotopic mass error units">Daltons</note>
    <note>The value for this parameter may be 'Daltons' or 'ppm': all other values are ignored</note>
    <note type="input" label="spectrum, parent monoisotopic mass error units">ppm</note>
        <note>The value for this parameter may be 'Daltons' or 'ppm': all other values are ignored</note>
    <note type="input" label="spectrum, fragment mass type">monoisotopic</note>
        <note>values are monoisotopic|average </note><note>spectrum conditioning parameters</note>
    <note type="input" label="spectrum, dynamic range">100.0</note>
        <note>The peaks read in are normalized so that the most intense peak
        is set to the dynamic range value. All peaks with values of less that
        1, using this normalization, are not used. This normalization has the
        overall effect of setting a threshold value for peak intensities.</note>
    <note type="input" label="spectrum, total peaks">50</note>
        <note>If this value is 0, it is ignored. If it is greater than zero (lets say 50),
        then the number of peaks in the spectrum with be limited to the 50 most intense
        peaks in the spectrum. X! tandem does not do any peak finding: it only
        limits the peaks used by this parameter, and the dynamic range parameter.</note>
    <note type="input" label="spectrum, maximum parent charge">4</note>
    <note type="input" label="spectrum, use noise suppression">yes</note>
    <note type="input" label="spectrum, minimum parent m+h">500.0</note>
    <note type="input" label="spectrum, minimum fragment mz">150.0</note>
    <note type="input" label="spectrum, minimum peaks">15</note>
    <note type="input" label="spectrum, threads">1</note>
    <note type="input" label="spectrum, sequence batch size">1000</note><note>residue modification parameters</note>
    <note type="input" label="residue, modification mass">57.022@C</note>
        <note>The format of this parameter is m@X, where m is the modfication
        mass in Daltons and X is the appropriate residue to modify. Lists of
        modifications are separated by commas. For example, to modify M and C
        with the addition of 16.0 Daltons, the parameter line would be
        +16.0@M,+16.0@C
        Positive and negative values are allowed.        </note>
    <note type="input" label="residue, potential modification mass"></note>
        <note>The format of this parameter is the same as the format
        for residue, modification mass (see above).</note>
    <note type="input" label="residue, potential modification motif"></note>
        <note>The format of this parameter is similar to residue, modification mass,
        with the addition of a modified PROSITE notation sequence motif specification.
        For example, a value of 80@[ST!]PX[KR] indicates a modification
        of either S or T when followed by P, and residue and the a K or an R.
        A value of 204@N!{P}[ST]{P} indicates a modification of N by 204, if it
        is NOT followed by a P, then either an S or a T, NOT followed by a P.
        Positive and negative values are allowed.        </note><note>protein parameters</note>
    <note type="input" label="protein, taxon">other mammals</note>
        <note>This value is interpreted using the information in taxonomy.xml.</note>
    <note type="input" label="protein, cleavage site">[RK]|{P}</note>
        <note>this setting corresponds to the enzyme trypsin. The first characters
        in brackets represent residues N-terminal to the bond - the '|' pipe -
        and the second set of characters represent residues C-terminal to the
        bond. The characters must be in square brackets (denoting that only
        these residues are allowed for a cleavage) or french brackets (denoting
        that these residues cannot be in that position). Use UPPERCASE characters.
        To denote cleavage at any residue, use [X]|[X] and reset the
        scoring, maximum missed cleavage site parameter (see below) to something like 50.        </note>
    <note type="input" label="protein, modified residue mass file"></note>
    <note type="input" label="protein, cleavage C-terminal mass change">+17.002735</note>
    <note type="input" label="protein, cleavage N-terminal mass change">+1.007825</note>
    <note type="input" label="protein, N-terminal residue modification mass">0.0</note>
    <note type="input" label="protein, C-terminal residue modification mass">0.0</note>
    <note type="input" label="protein, homolog management">no</note>
        <note>if yes, an upper limit is set on the number of homologues kept for a particular spectrum</note><note>model refinement parameters</note>
    <note type="input" label="refine">yes</note>
    <note type="input" label="refine, modification mass"></note>
    <note type="input" label="refine, sequence path"></note>
    <note type="input" label="refine, tic percent">20</note>
    <note type="input" label="refine, spectrum synthesis">yes</note>
    <note type="input" label="refine, maximum valid expectation value">0.1</note>
    <note type="input" label="refine, potential N-terminus modifications"></note>
    <note type="input" label="refine, potential C-terminus modifications"></note>
    <note type="input" label="refine, unanticipated cleavage">yes</note>
    <note type="input" label="refine, potential modification mass"></note>
    <note type="input" label="refine, point mutations">no</note>
    <note type="input" label="refine, use potential modifications for full refinement">no</note>
    <note type="input" label="refine, point mutations">no</note>
    <note type="input" label="refine, potential modification motif"></note>
    <note>The format of this parameter is similar to residue, modification mass,
        with the addition of a modified PROSITE notation sequence motif specification.
        For example, a value of 80@[ST!]PX[KR] indicates a modification
        of either S or T when followed by P, and residue and the a K or an R.
        A value of 204@N!{P}[ST]{P} indicates a modification of N by 204, if it
        is NOT followed by a P, then either an S or a T, NOT followed by a P.
        Positive and negative values are allowed.        </note><note>scoring parameters</note>
    <note type="input" label="scoring, minimum ion count">4</note>
    <note type="input" label="scoring, maximum missed cleavage sites">1</note>
    <note type="input" label="scoring, x ions">no</note>
    <note type="input" label="scoring, y ions">yes</note>
    <note type="input" label="scoring, z ions">no</note>
    <note type="input" label="scoring, a ions">no</note>
    <note type="input" label="scoring, b ions">yes</note>
    <note type="input" label="scoring, c ions">no</note>
    <note type="input" label="scoring, cyclic permutation">no</note>
        <note>if yes, cyclic peptide sequence permutation is used to pad the scoring histograms</note>
    <note type="input" label="scoring, include reverse">no</note>
        <note>if yes, then reversed sequences are searched at the same time as forward sequences</note>
    <note type="input" label="scoring, cyclic permutation">no</note>
    <note type="input" label="scoring, include reverse">no</note><note>output parameters</note>
    <note type="input" label="output, log path"></note>
    <note type="input" label="output, message">testing 1 2 3</note>
    <note type="input" label="output, one sequence copy">no</note>
    <note type="input" label="output, sequence path"></note>
   <note type="input" label="output, path">output.xml</note>
    <note type="input" label="output, sort results by">protein</note>
        <note>values = protein|spectrum (spectrum is the default)</note>
    <note type="input" label="output, path hashing">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, xsl path">tandem-style.xsl</note>
    <note type="input" label="output, parameters">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, performance">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, spectra">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, histograms">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, proteins">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, sequences">yes</note>
        <note>values = yes|no</note>
    <note type="input" label="output, one sequence copy">no</note>
        <note>values = yes|no, set to yes to produce only one copy of each protein sequence in the output xml</note>
    <note type="input" label="output, results">valid</note>
        <note>values = all|valid|stochastic</note>
    <note type="input" label="output, maximum valid expectation value">0.1</note>
        <note>value is used in the valid|stochastic setting of output, results</note>
    <note type="input" label="output, histogram column width">30</note>
        <note>values any integer greater than 0. Setting this to '1' makes cutting and pasting histograms
        into spread sheet programs easier.</note><note type="description">ADDITIONAL EXPLANATIONS</note>
    <note type="description">Each one of the parameters for X! tandem is entered as a labeled note
            node. In the current version of X!, keep those note nodes
            on a single line.    </note>
    <note type="description">The presence of the type 'input' is necessary if a note is to be considered
            an input parameter.    </note>
    <note type="description">Any of the parameters that are paths to files may require alteration for a
            particular installation. Full path names usually cause the least trouble,
            but there is no reason not to use relative path names, if that is the
            most convenient.    </note>
    <note type="description">Any parameter values set in the 'list path, default parameters' file are
            reset by entries in the normal input file, if they are present. Otherwise,
            the default set is used.    </note>
    <note type="description">The 'list path, taxonomy information' file must exist.        </note>
    <note type="description">The directory containing the 'output, path' file must exist: it will not be created.        </note>
    <note type="description">The 'output, xsl path' is optional: it is only of use if a good XSLT style sheet exists.        </note></bioml>

4. 结果

软件得到的结果也是xml文件，文件名会自动加上运行的日期，如output.2020_07_03_15_56_03.t.xml。对格式需要有一定了解才能从中有效的提取信息，关于输出的文件格式解释，参考这份文档：https://www.thegpm.org/docs/X_series_output_form.pdf

这里从我跑过的一个项目中截取一小部分结果来做说明：

<?xml version="1.0"?><?xml-stylesheet type="text/xsl" href="tandem-style.xsl"?><bioml xmlns:GAML="http://www.bioml.com/gaml/" label="models from '/path/test.mgf'"><group id="373508" mh="1842.861248" z="3" rt="6989.4672" expect="8.7e-12" label="[denovogenes]_384740" type="model" sumI="7.12" maxI="1.08216e+06" fI="10821.6" act="0" ><protein expect="-196.9" id="373508.1" uid="2902844" label="[denovogenes]_384740" sumI="9.09" ><note label="description">[denovogenes]_384740</note><file type="peptide" URL="/path/test.fa"/><peptide start="1" end="434">
    MSIITDVYAR EVLDSRGNPT LEVEVYTESG AFGRGMVPSG ASTGEHEAVE
    LRDGDKARYG GLGTQKAVDN VNNVIAEHII GFDVRDQQGI DRAMIALDGT
    PNKGKLGANA ILGVSIAVAR AAADYLEVPL YSYLGGFNTK VLPTPMMNII
    NGGSHSDAPI AFQEFMIVPA GAPTFKEALR WGAEIFHALK KILKERGLET
    AVGDEGGFAP RFDGTEDGVE TIIKAIEAAG YVPGKDVFIG FDCASSEFYD
    AERKVYDYTK FEGEGAAVRT AAEQIDYLEE LVNKYPIITI EDGMDENDWD
    GWKALTERLG GKVQLVGDDF FVTNTAYLEK GIAEHAANSI LIKVNQIGTL
    TETFDAIEMA KEAGYTAVVS HRSGETEDST IADIAVATNA GQIKTGSLSR
    TDRIAKYIQL LRIEEQLGEV AEYRGLKSFY NLKK<domain id="373508.1.1" start="35" end="52" expect="8.7e-12" mh="1842.8650" delta="-0.0038" hyperscore="83.7" nextscore="45.9" y_score="9.5" y_ions="23" b_score="10.1" b_ions="6" pre="AFGR" post="DGDK" seq="GMVPSGASTGEHEAVELR" missed_cleavages="0"><aa type="M" at="36" modified="15.99492" /></domain></peptide>......

假设我只提取蛋白ID，delta，domain起始和终止，期望和序列等信息。xml格式在各种编程语言中都有解析包，自己写脚本提取后可得到如下形式，再进行后续处理（如宏蛋白中优化数据库，将鉴定的蛋白ID匹配回原始数据库，从而达到优化数据库的目的）。

5. FAQ

最常见的FAQ官网已经给出，基本上这些也够了。
https://www.thegpm.org/TANDEM/tandem_install_faq.html
主要是理解三个输入的配置文件input.xml，taxonomy.xml，default_input.xml以及知道它们干什么的。

蛋白质组学鉴定定量系列软件总结：
【1】蛋白鉴定软件之X!Tandem
【2】蛋白鉴定软件之Comet
【3】蛋白鉴定软件之Mascot
【4】蛋白质组学鉴定软件之MSGFPlus
【5】蛋白质组学鉴定定量软件之PD
【6】蛋白质组学鉴定定量软件之MaxQuant