parse xml proteinpredict files

Question

[Biojl]

Hi,

I have several thousands of xml files from https://www.predictprotein.org/ calculations for different proteins. I was wondering if anyone knows a package to parse that information in python or R in order to be able to perform some calculations easily.

I am mostly interested in the secondary structure information. Obtaining the relative coordinates of every feature and convert it to a bed file. An example file here: https://raw.githubusercontent.com/gyachdav/pp-results/master/examples/ADRB2_HUMAN.xml

[Pierre Lindenbaum]

please provide a sample of XML. What kind of information do you want to retrieve ? Most a of the time, a simple XSL stylesheet to the job if it is a simple query .

[Biojl]

Hi, I have updated the question with the relevant information, including an example file. I am not familiar with XSL stylesheets but I'll dig into it.

[Pierre Lindenbaum]

. Obtaining the relative coordinates of every feature and convert it to a bed file

That's not clear to me. give me a few lines for an example please.

[Biojl]

I'm interested in the info inside this feature: <featuretypegroup type="secondary structures">...</featuretypegroup>

Some example lines inside those tags:

<feature type="helix" soTermId="SO:0001114">
<location>
<begin position="89"/>
<end position="108"/>
</location>
</feature>
<feature type="strand" soTermId="SO:0001111">
<location>
<begin position="109"/>
<end position="113"/>
</location>

The idea would be to get:

ENSP0001, helix, 89, 108 \n
ENSP0001, strand, 109, 113

[Pierre Lindenbaum]

I cannot find ENSP0001 in your example.

[Biojl]

Is not there, is a made up protein ID. The protein ID is usually in the file name.

Answer 1

[Pierre Lindenbaum]

using the following xslt stylesheet:

$ xsltproc biostars293423.xsl input.xml

ADRB2_HUMAN protein binding region  26  28  
ADRB2_HUMAN protein binding region  147 147 
ADRB2_HUMAN protein binding region  179 180 
ADRB2_HUMAN protein binding region  236 236 
ADRB2_HUMAN protein binding region  243 243 
ADRB2_HUMAN protein binding region  248 252 
ADRB2_HUMAN protein binding region  343 347 
ADRB2_HUMAN disordered region   9   10  
ADRB2_HUMAN disordered region   19  21  
ADRB2_HUMAN disordered region   389 389 
ADRB2_HUMAN disordered region   406 406 
ADRB2_HUMAN disordered region   1   30  
ADRB2_HUMAN disordered region   61  65  
ADRB2_HUMAN disordered region   140 148 
ADRB2_HUMAN disordered region   175 183 
ADRB2_HUMAN disordered region   186 196 
ADRB2_HUMAN disordered region   228 270 
ADRB2_HUMAN disordered region   299 305 
ADRB2_HUMAN disordered region   330 331 
ADRB2_HUMAN disordered region   334 334 
ADRB2_HUMAN disordered region   343 413 
ADRB2_HUMAN disordered region   21  25  
ADRB2_HUMAN disordered region   228 233 
ADRB2_HUMAN disordered region   235 235 
ADRB2_HUMAN disordered region   359 359 
ADRB2_HUMAN disordered region   366 376 
ADRB2_HUMAN disordered region   394 401 
ADRB2_HUMAN disordered region   404 405 
ADRB2_HUMAN disordered region   1   2   
ADRB2_HUMAN disordered region   356 359 
ADRB2_HUMAN disordered region   363 376 
ADRB2_HUMAN disordered region   379 379 
ADRB2_HUMAN disordered region   381 411 
ADRB2_HUMAN strand  31  38  
ADRB2_HUMAN helix   39  41  
ADRB2_HUMAN strand  42  48  
ADRB2_HUMAN strand  52  60  
ADRB2_HUMAN helix   68  86  
ADRB2_HUMAN helix   89  97  
ADRB2_HUMAN helix   103 115 
ADRB2_HUMAN helix   117 122 
ADRB2_HUMAN strand  123 128 
ADRB2_HUMAN strand  133 136 
ADRB2_HUMAN helix   147 151 
ADRB2_HUMAN strand  152 154 
ADRB2_HUMAN helix   155 165 
ADRB2_HUMAN strand  170 175 
ADRB2_HUMAN strand  197 205 
ADRB2_HUMAN helix   209 231 
ADRB2_HUMAN helix   262 272 
ADRB2_HUMAN strand  273 285 
ADRB2_HUMAN helix   288 298 
ADRB2_HUMAN helix   305 318 
ADRB2_HUMAN strand  324 327 
ADRB2_HUMAN helix   330 342

[Biojl]

Wow, definitely is worth it to learn how to use XSL!! Just one question about the code, in which part you select the info for the first column (ADRB2_HUMAN)? I don't understand this language yet.

[Pierre Lindenbaum]

n which part you select the info for the first column (ADRB2_HUMAN)

<xsl:value-of select="/p:predictprotein/p:entry/p:accession"/>