Convert blastx XML to Fasta w/ NTs, not Proteins?

[Blaze9]

Hey,

I have a blastx xml output. I want to convert this blastx back into a fasta file for downstream analysis. However, the xml contains only protein sequences since it's a blastx output. I have the original fasta I sent into blast. Is there a way I can "merge" these two files together? I want to take the annotation from the xml and append it to the original fasta's sequences.

Is this possible via command line? Right now I'm doing this by importing my fasta and XML into blast2go and exporting the fasta w/ the new header information, but blast2go gets really slow with very large datasets.

[lindenb]

an idea (not tested) you could create a awk script from the xml using XSLT

<?xml version='1.0'  encoding="UTF-8" ?>
<xsl:stylesheet  xmlns:xsl="http://www.w3.org/1999/XSL/Transform"  version='1.0'>

<xsl:output method="text" />


<xsl:template match="/">
<xsl:apply-templates select="BlastOutput/BlastOutput_iterations/Iteration"/>
</xsl:template>


<xsl:template match="Iteration">
<xsl:text>$1 == "&gt;</xsl:text><xsl:value-of select="Iteration_query-def/text()"/> <xsl:text>"	{ </xsl:text>
<xsl:for-each select="Iteration_hits/Hit/Hit_hsps/Hsp">printf("%s [%s-%s]\n%s\n",$1,<xsl:value-of select="Hsp_query-from/text()"/>,<xsl:value-of select="Hsp_query-to/text()"/>,substr($2,<xsl:value-of select="Hsp_query-from/text()"/>,<xsl:value-of select="Hsp_query-to/text()"/>));</xsl:for-each>
<xsl:text> next;}
</xsl:text>
</xsl:template>
</xsl:stylesheet>

xsltproc stylesheet.xsl input.blastx.xml  > my.awk

then, linearise the input fasta and use this awk script to extract the substring

awk '/>/ {printf("\n%s\t",$0);next;} {printf("%s",$0);} END {printf("\n");}' query.fasta  |\
  awk -F '\t' -f my.awk

if the input fasta is too large, you could index the query.fasta with makeblastdb and call blastdbcmd with 'range' instead of awk