Star for Soybean RNA-seq aligment

Question

[valopes]

Hi all

It is my first time working with STAR and I need to generate the genome index. However I have just the gff3 file intead of gtf. My gff3 file shows gene, mRNA, CDS, five_prime_UTR and three_prime_UTR annotation. Plus, everytime I try to run it, I am exceeding the RAM usage.

I am try this command:

STAR --runMode genomeGenerate     \
          --genomeDir ./genomeDir       \
          --genomeFastaFiles genomeDir/Gmax_275_v2.0.fa  \
          --runThreadN 8      \
          --limitGenomeGenerateRAM 32000000000 \
          --genomeChrBinNbits 10
          --sjdbGTFfile genomeDir/Gmax_275_Wm82.a2.v1.gene.gff3   \
          --sjdbGTFtagExonParentGene ID  \
          --sjdbGTFfeatureExon CDS  \
          --sjdbOverhang 99

Am I doing it right? It looks it is not working for me.

Thank you all

[Devon Ryan]

Am I correct in assuming that the soy bean genome is really really large? How much RAM do you have?

[GenoMax]

The soybean genome has 20 chromosomes and an estimated size of 1,115 Mbp.

Answer 1

[Michael]

Despite the name, the sjdbGTFfile parameter accepts GFF3, we have tried this without problems. Well, RAM usage was high with the default settings, if I remember correctly. We needed about 300GB for generating the index for our 670Mb/33k scaffolds genome. Generating an index for the Salmon genome required almost 1TB. There are parameter settings that should allow for much lower RAM usage. See: http://seqanswers.com/forums/showthread.php?t=27470