Error in BLAT Program

Question

[Ram]

Hi all,

As I am trying to use BLAT to remap all reads supporting a variant to the genome.

But I am always getting below error with mentioned command.

./blat -infile filtered_intron.txt -genomefile hg19.2bit -outfile filtered_intron_blat.txt

needMem: trying to allocate 2656204096 bytes (limit: 500000000)

Any help!

Thanks.

Answer 1

[GenoMax]

It appears that blat is trying to use 2.6G when only 0.5G is available. This requirement would be in-line with what human genome index will require in terms of RAM.

[Ram]

So how is it possible to change over to more memory .

As I also made some suggested changes in memalloc.c. But it didn't helped.

Thanks

[GenoMax]

Are you running this in a VM? How much memory do you have? It appears that the OS is reporting that there is a limit of 0.5G (per process?)

[WouterDeCoster]

You have enough memory on your machine? Assuming yes: Using 'ulimit -a' in a terminal will show you the restrictions on memory usage and you may be able to increase those. If not, you may want to talk to your sys admin.

Answer 2

[Ram]

Hey Thanks all,

I solved the part with memory problem. But with same command I am getting different error :

couldn't open chr1 , no such file or directory

Any idea ?

Thanks

[GenoMax]

Blat probably needs the fasta files when 2bit indexes are used for alignment. You should be able to regenerate the fasta files with the twoBitToFa command.

[Ram]

Hey @ genomax2, Yes with the fasta file also it's showing the same error.

[GenoMax]

The order of files is important here.

$ ./blat hg19.2bit filtered_intron.txt filtered_intron_blat.txt

usage: blat database query [-ooc=11.ooc] output.psl where:
database and query are each either a .fa, .nib or .2bit file, or a list of these files with one file name per line.

by default blat will produce PSL format output (tab delimited) if you need the output to be in some other format look at the options for -out command modifier.

   -out=type      Controls output file format.  Type is one of:
                     psl - Default.  Tab-separated format, no sequence
                     pslx - Tab-separated format with sequence
                     axt - blastz-associated axt format
                     maf - multiz-associated maf format
                     sim4 - similar to sim4 format
                     wublast - similar to wublast format
                     blast - similar to NCBI blast format
                     blast8- NCBI blast tabular format
                     blast9 - NCBI blast tabular format with comments

You can see help for blat by running the command without any options on the command line.

[Ram]

Hey Thanks for your reply,

But still the same error after changing the position also :

./blat ucsc.hg19.fasta filtered_intron.txt -out=psl filtered_intron_blat.psl Loaded 3137161264 letters in 93 sequences Couldn't open chr1 , No such file or directory

Thanks

[GenoMax]

Hmm. That is odd. What does your filtered_intron.txt file look like? I have a feeling that it is not in fasta (but probably BED) format.
$ head -5 filtered_intron.txt

[Ram]

Hey thanks for reply. It looks like this

chr1:14560-14561 G chr1:14561-14562 A chr1:14562-14563

But now its giving another error :

Loaded 16448246 letters in 1 sequences Invalid fasta format: sequence size == 0 for element chr1:10366140-1 Searched 701839 bases in 701840 sequences

Thanks

[GenoMax]

Looks like your input file has variants/SNP's listed for specific positions? This file is NOT going to work with blat (which requires a simple fasta format file at a minimum). Are these SNP's from introns?
Can you provide additional clarification as to what you are trying to do?

[Ram]

Yes these are SNPs from introns. I am using BLAT to remap all reads supporting a variant to the genome. For it I am using SNPiR pipeline to know RNA-editing sites.