Entering edit mode
15 months ago
camillab.
▴
160
Hi,
I am trying to build the index for the STAR alignment and it basically doesn't work as it does not progress at all. I have a M1 Mac and I have enough memory, is my computer the problem? I tried yesterday my code on my old Mac and run easily (I stopped because it was just a test).
My code is:
/Users/camilla/Desktop/FASTQ/STAR-2.7.11a/bin/MacOSX_x86_64/STAR runMode genomeGenerate --genomeDir /Users/camilla/Desktop/diofa/GENOMEDIR/index --genomeFastaFiles /Users/camilla/Desktop/diofa/genome/genome/GRCh38.primary_assembly.genome.fa --sjdbGTFfile /Users/camilla/Desktop/diofa/genome/genome/gencode.v44.annotation.gtf
And what I get is:
STAR version: 2.7.11a compiled: :/Users/distiller/project/STARcompile/source
Sep 20 09:53:05 ..... started STAR run
But it doesn't do anything else except creating a folder.
Thank you
Camilla
Please use STAR's RAM usage options whenever you use STAR.
--limitGenomeGenerateRAM
is the relevant RAM usage parameter here.Do you think it will make it work on my Mac that has less than 30G RAM?
It gives it a better chance. It can't bypass what STAR needs but it can limit what STAR gets access to, ensuring a safe failure that does not impact other programs that also need your RAM.
thank you! I am following GenoMax suggestion and I am trying with
salmon
but it is stuck as well. it's not making any progress.....STAR is known to require upwards of 30G of RAM with human genome (to give you size reference) . What is the size of your reference? Perhaps you simply need to be patient. Creating indexes can take a while, up to a half hour.
I have 16G RAM but my old computer (that it runs- it takes ages) doesn't have 30G of RAM for sure. so shall I just wait?
Open a new terminal. Run
top
to see if the STAR process is actively running (understate
column) . If it is wait and watch. If the process runs out of memory you will get an error.Thanks, I did it but It's not even in the list. is it normal?
Do you get the system prompt back right away? If you do then the process is failing.
Is there a transcriptome available? Perhaps you could use
salmon
which will have much lower hardware requirements instead.Even if you managed to create STAR index, you don't seem to have sufficient memory to align reads to the genome.
STAR’s high mapping speed is traded off against RAM usage: STAR requires about 27 GB of RAM for aligning to the human genome.
https://academic.oup.com/bioinformatics/article/29/1/15/272537