Making a subset of spectra from an mzXML file in Python

[Ben Temperton]

Hi there,

I am trying to pull out a subset of data from an mzXML file to run against a database using MSGF-db (for instance, to re-run any non matching spectra against the database searching for phosphorylation). To generate the subset I am currently using:

import lxml.etree as le

SASHIMI_NAMESPACE = 'http://sashimi.sourceforge.net/schema_revision/mzXML_3.1'

def makeHQSpectraFile(spectraFile, spectraList, outputFile):

    """Takes a spectra file, a list of scan ids to include and an output file as parameters"""

    with open(spectraFile,'r') as f:

        doc=le.parse(f)

        root = doc.getroot()

        for elem in doc.xpath('/t:mzXML/t:msRun/t:scan', namespaces={'t' : SASHIMI_NAMESPACE}):

            if not elem.attrib['num'] in spectraList:

                parent=elem.getparent()

                parent.remove(elem)

        for elem in doc.xpath('/t:mzXML/t:index/t:offset', namespaces={'t' : SASHIMI_NAMESPACE}):

            if not elem.attrib['id'] in spectraList:

                parent=elem.getparent()

                parent.remove(elem)

    handle = open(outputFile, 'wb')

    handle.write(le.tostring(doc) + '\n')

    handle.close()

However, when I run MSGF-db on the new file it throws a:

Reading spectra...

javax.xml.stream.XMLStreamException: ParseError at [row,col]:[1,1]

Message: Premature end of file.

at com.sun.org.apache.xerces.internal.impl.XMLStreamReaderImpl.next(Unknown Source)

at org.systemsbiology.jrap.stax.IndexParser.parseIndexes(IndexParser.java:176)

at org.systemsbiology.jrap.stax.MSXMLParser.randomInits(MSXMLParser.java:117)

at org.systemsbiology.jrap.stax.MSXMLParser.<init>(MSXMLParser.java:134)

at parser.MzXMLSpectraMap.<init>(MzXMLSpectraMap.java:39)

at parser.MzXMLSpectraIterator.<init>(MzXMLSpectraIterator.java:36)

at parser.MzXMLSpectraIterator.<init>(MzXMLSpectraIterator.java:26)

at ui.MSGFDB.runMSGFDB(MSGFDB.java:269)

at ui.MSGFDB.runMSGFDB(MSGFDB.java:106)

at ui.MSGFDB.main(MSGFDB.java:82)

Whilst the original (non-parsed version) works fine. I can't get the mzXMLValidator to work on our systems (see post here https://groups.google.com/d/msg/spctools-discuss/bAxu-In-ju4/z9_g3mdWSFcJ), so I was wondering if anyone else had ever encountered a similar issue and had any tips.

Many thanks,

Ben

[Jimmy Eng]

Ben,

Have you done anything special to handle the scan numbers (which
presumably are not consecutive anymore starting from scan 1) and the
scan index? If not, address those and re-test or find out if those
are important for MSGF-db.

> --
> You received this message because you are subscribed to the Google Groups
> "spctools-discuss" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/spctools-discuss/-/psC3ABG8sNcJ.
> To post to this group, send email to spctools...@googlegroups.com.
> To unsubscribe from this group, send email to
> spctools-discu...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/spctools-discuss?hl=en.

[Ben Temperton]

Hi Jimmy,

I have updated the scan numbers using a new method (attached), but I am stuck at changing the offset values for each of the scans in the /index/offset elements. I think this is what is causing the 'premature end of file' error - by not changing the offsets, MSGF-db can't find the spectra.

Any ideas how I can regenerate the index once the poor-quality spectra have been removed?

[Brian Pratt]

The index is technically optional, you should be able to just skip it
in your output.

("technically" and "should", as some parsers are brittle and will fail
without it - good luck!)

> --
> You received this message because you are subscribed to the Google Groups
> "spctools-discuss" group.
> To view this discussion on the web visit

> https://groups.google.com/d/msg/spctools-discuss/-/egyhnHrt7HsJ.