easyRNAseq annotation file

[sylvian]

Hi,

Apparently I need to use a gtf or gff3 file for annotation with easyRNAseq now when processing hg19 bamfiles now - as method biomaRt uses GRCh39 and I get an error. So far so bad.

However, I seem not to be able to download a gtf or gff3 file for hg19 that easyRNAseq would accept:

> DGE<-simpleRNASeq(bamFiles=getBamFileList(bamfiles), param=param, nnodes=4, verbose=T, override=T)


...

==========================
Processing the annotation
==========================
Validating the annotation source
Read 993 records
Error in .validate(obj, verbose = verbose) :
The provided gff3 contains no annotation of type 'mRNA' and/or 'exon' in the first 1000 lines.


I've tried files the following annotation files but none seems to work:

ref_GRCh37.p13_top_level.gff3.gz
ref_GRCh37.p13_scaffolds.gff3.gz
gencode.v19.annotation.gtf
gencode.v19.annotation.gff3
Homo_sapiens.GRCh37.70.gtf

will try the Illumina iGenome next. What am I missing here?!?

[iandry]

Hi,

I have the same problem with mm10 annotations...

did someone found a solution?

[mastal]

Have you had a look at section 7.3 of the easyRNASeq manual,

7.3 Dealing with annotation inconsistencies

it might help.

http://bioconductor.org/packages/release/bioc/vignettes/easyRNASeq/inst/doc/easyRNASeq.pdf

Problems could be due to the fact that UCSC and Ensembl/BioMart use different names for the chromosomes, for the gtf to work you need to have the same chromosome names in the bam file as in the gtf.

What error message are you getting?

[iandry]

I already checked for the chromosomes names

with gtf files from Ensembl or UCSC, I had this error message

==========================
Processing the annotation
==========================
Validating the annotation source
Read 112 records
Error in .validate(obj, verbose = verbose) :
Your gtf file: /ep10/disks/SANB5/data/genome/mouse_mm9_bowtie2/gtf/Mus_musculus.GRCm38.83.changed.gtf does not contain all the required fields: gene_id, transcript_id, exon_id.

with the gff3, I had this

==========================
Processing the annotation
==========================
Validating the annotation source
Read 874 records
Error in .validate(obj, verbose = verbose) :
The provided gff3 contains no annotation of type 'mRNA' and/or 'exon' in the first 1000 lines.

And with annotation using biomaRt, I had

Error in simpleRNASeq(bamFiles = bam, param = rnaSeqParam) :
There is no common genomic references between your BAM
files and the provided annotation. Fix one or the other.

[iandry]

Still don't work with gtf or gff file, but I found the solution with biomart annotation.

Now I have this error message :

==========================
Creating the count table
==========================
Using 40 CPU cores in parallel
Error in unserialize(node$con) : error reading from connection

I tried with 10 cores, i had the same error message.

Is the parallelization really working ?