Running a tools depending on old version of Picard that cause problem

Hello,

I’m trying to make the following tool run (http://www.appliedbioinformatics.com.au/index.php/SGSGeneLoss), but I cant suceed…

I think that the problem come from an old version of Picard-tools (1,89) which the tool depends on. I’m unable to make this version of Picard tools to run. I tried to download precompiled files from sourceforce (https://sourceforge.net/projects/picard/files/picard-tools/1.89/) and I also built it myself using the github version, but nothing works, when I try to run picard.jar/sam.jar; which are the two dependencys of the tool, I always get an error message, which is different depending of the version of java I’m using.

For example :

Java 6 (openjdk version "1.6.0-113") from zulu:
/usr/lib/jvm/zulu-6-amd64/jre/bin/java -jar picard-1.89.jar
Exception in thread "main" java.lang.NullPointerException
at sun.launcher.LauncherHelper.getMainClassFromJar(LauncherHelper.java:353)
at sun.launcher.LauncherHelper.checkAndLoadMain(LauncherHelper.java:402)

java 6 (1.6.0_45) from oracle :
/usr/lib/jvm/oracle-java6-jdk-amd64/bin/java -jar picard-1.89.jar
Failed to load Main-Class manifest attribute from
picard-1.89.jar

java 8 (openjdk version "1.8.0_222")
java -jar picard-1.89.jar
no main manifest atribute in picard-1.89.jar

This problem is specific to the two tools I need (picard.jar and sam.jar). The others tools included in picard-tools 1.89 seems to work...

Do you have an idea of where the problem could be from ? I’m lost and It seems like I reached the limits of my skills…

Thanks in advance.

sgsgeneloss sgs picard software error pav

• 1,821 views • link

written 6.9 years ago by Alexso •

1 answer

This is not how the old version of picard worked. It should be something like, say for SortSam

java -cp picard-1.89.jar:sam-1.89.jar:variant-1.89.jar:SortSam.jar  net.sf.picard.sam.SortSam

furthermore, your version of java is really old , I don't think it would work. Try to upgrade to java 1.8

• 0 views • link

written 6.9 years ago by Pierre Lindenbaum

Thanks you for this answer.

Indeed, if I try to run Picard like this, it works… It’s depressing ! So the problem is not from Picard…

I Installed old versions of Java in order to make some test with Picard. I already have 1.8 Installed.

I only get this error message when I try to use SGSGeneLoss :

java.lang.IndexOutOfBoundsException: Index: 0, Size: 0
        at java.util.ArrayList.rangeCheck(ArrayList.java:657)
        at java.util.ArrayList.get(ArrayList.java:433)
        at ExCovUtils.mergeExons(ExCovUtils.java:477)
        at ExCovUtils.gatherGenesGff3All(ExCovUtils.java:84)
        at ExCovUtils.getAllGenesGff3(ExCovUtils.java:457)
        at ExonCoverage5.main(ExonCoverage5.java:259)
Exception in thread "main" java.lang.NullPointerException
        at ExCovUtils.genesListToMap(ExCovUtils.java:463)
        at ExonCoverage5.main(ExonCoverage5.java:265)

So the problem is probably comming from one of my input file…

• 0 views • link

written 6.9 years ago by Alexso •

o the problem is not from Picard…

or it's a bug in picard that has been fixed in the new versions...

• 0 views • link

written 6.9 years ago by Pierre Lindenbaum

They say that the tool has been tested on 1.89, so it's supposed to work (I hope so !)

If I look to the lines indicated by the error message above, they all refer to the gff3 parsing part of the program. It's as if there was a problem parsing it...

As I was wrotting this, I tried to create a very small gff with only one entry. And with this one it seems to run. I will have fun testing things like that !

Edit : It also work with two genes... It looks good !

Edit 2 : I found out what was the problem... It was some non-coding RNA that had no "exon" part in the gff...

• 0 views • link

written 6.9 years ago by Alexso •