problem with Scripture.jar for RNA-seq

[lj907]

Hi all,

I'm trying to run Scripture.jar which I have downloaded from the Broad institute website: http://www.broadinstitute.org/softwa...ipture/?q=home
I'm on a Mac and I have java 6 and 7 (I know that by checking /utilites/java preferences). For the sake of practice, I'm trying to run the walkthrough exercise in the tutorial for this software (http://www.broadinstitute.org/softwa...hrough_example)... in step 7 you need to run scripture with the sam file you have generated in the previous steps and afterwards downloading the mouse size file and the mouse chr19 sequence... this is the command line I'm using:
>>java –jar scripture.jar –alignment all_alignments.sorted.sam –out chr19.scriptureESTest.segments –sizeFile mm9.sizes –chr chr19 –chrSequence chr19.fa
However, I'm getting this response:
Exception in thread "main" java.lang.IllegalArgumentException: Argument alignment is mandatory
Does anyone know what the problem is....
I appreciate your support
lj907

[sdriscoll]

i think you need to download one of the beta versions. i had this same problem. after i got the beta version i made it a little further through the walkthrough only for it to crash with a null pointer exception near the end of the assembly process. so i'm stuck as well.

http://www.broadinstitute.org/software/scripture/?q=download

[lj907]

Thanks sdriscoll. I have now shifted to using htseq-count followed by DESeq for differential expression. The Cufflinks path is probably better for novel transcript discovery than differential expression. Scripture was just another option which I explored to some extent, but never could get it to work, and there seems to be no support for it what so ever. The author shouldn't expect many citations.

[sdriscoll]

i'd also recommend trying out eXpress or RSEM for quantification. it's a valid argument that "correct" gene level expression is dependent upon proper disambiguation of aligned reads in a genome with alternative splicing. both eXpress and RSEM rely on the EM algorithm. eXpress has been optimized so it works much quicker and you can even pipe alignments right into it from bowtie. both of those tools work by aligning to a transcriptome allowing all alignments with very little limitations from the aligner aside from mismatches in the seed of each read. the ability of these tools to determine if an alignment is correct or not have surpassed those in the aligner so they suggest allowing their software to determine alignment. also both of these tools can perform reliable isoform level expression meaning you can still do isoform level DE and look for changes in the relative levels of isoform expressions within loci across samples.

i've been testing eXpress today and i really like it. the output is very informative as well - more informative than the output of RSEM though the output of RSEM is a little cleaner and they organize the results into gene level expressions for you. with eXpress you have to sum expressions & counts by gene name after the fact. also RSEM performs the alignments for you while eXpress requires you do that yourself. anyway - they both have sites - check 'em out!