I have a bunch of FASTQ files (10 samples). I try to load them in R for mapping them but I cannot figure out how to do For example, Dseq2 requires bat files and not fastq, so I cannot find the way how to import fastq and analysis
I tried to use Rsubread 1.26.1 but for example after I do the following command I get this error ERROR: the output directory is not writable, but the index builder needs to create temporary files in the current directory. Please change the working directory and rerun the index builder.
I do setwd() to a good place but still I get error
Can someone comment how to start the analysis for Fastq files ?
Why are you trying to use R to do mapping? Isn't that better done on the command line in unix?
@genomax it does not matter for me I thought it is easy in R :-) but the main reason is because i want to do expression analysis in R, so I thought R is the best way to go . can you give me a solution ? I will try to do it or give me the direction.
What aligner(s) do you have access to/have installed on your system? STAR, bbmap, HISAT2 can all be options on the command line. I suppose
Rsubreadshould work too in R but I don't have any experience with it.
@genomax I have Mac, I believe I can install any of them, which one is more stable ? or recomened more to be used ? because I believe one should have a reference for alignment so I don't know how any of them works. Please advise
STAR is quite often used and pretty fast, but requires a lot of RAM (25-30GB) for big genomes. Are you working on human data?
@WouterDeCoster yes it is human data but I only have 10 samples , do you think that I need such a Ram?
Yes, that's what you need for loading the index while aligning a single sample. I don't know about your computer specs, but perhaps HISAT2 can be used then. Or do you have access to a cluster/server for bigger jobs?
Note that you can also use more lightweight pseudo-alignment approaches such as salmon, which nicely work with DESeq2 using tximport, see for example here. These shouldn't be a problem on "slightly less strong" hardware.
Yep. STAR/bbmap will need ~30G of RAM for human genome.
@genomax I have a mac with 32 GB of ram or a windows7 cluster with 64 RAM. let me know which one should i use?
Try STAR out on your Mac and see if it works. As long as you do one sample at a time it may. You could download pre-made STAR indexes from here.
@genomax Thanks, I am having problem with this STAR , I downloaded it , then I cd to the source and then I get error like this make STARforMacStatic CXX=/usr/bin/gcc /usr/bin/gcc -c -O3 -pipe -std=c++11 -Wall -Wextra -fopenmp -D'COMPILATION_TIME_PLACE="Mon Feb 12 16:03:45 CST 2018 :/Users/admin/Desktop/STAR-master/source"' -D'COMPILE_FOR_MAC' SharedMemory.cpp clang: error: unsupported option '-fopenmp' make: * [SharedMemory.o] Error 1
Get pre-compiled binary for macOS here.
@genomax I I got both of them, can you please now tell me how to use it ? should their format be .dms right?
STAR is the executable. You may need to do
sudo chmod a+x /path_to/STARto make it executable via terminal.
@genomax which one of these I should download http://labshare.cshl.edu/shares/gingeraslab/www-data/dobin/STAR/STARgenomes/ENSEMBL/homo_sapiens/ENSEMBL.homo_sapiens.release-83/
You have to download all of them. These file together form the index set.
Human RNA-seq using bwa? I wouldn't call that the next best bet tbh.
I don't pay attention to tags on posts since they don't always go with contents of posts :-)
I will edit my comment in post above.
Yes, tags are trouble, but this time there was additional evidence ;)