ENCODE long RNA-seq remapped

[alexdobin]

Dear All,

there have been multiple questions about the ENCODE RNA-seq alignments on the UCSC portal. These alignments had been generated by a 3-year old version of STAR and use some non-conventional formatting (e.g. they are not compatible with Cufflinks).

To bring this data up-to-date, I have remapped it using the latest version of STAR. The new alignments use conventional formatting and should be compatible with most downstream software. Importantly, annotations are used to improve the mapping accuracy. The BAMs for all of the ENCODE phase 2 (2008-2012) long RNA-seq data can be downloaded here:

http://it-collab01.cshl.edu/shares/gingeraslab/www-data/dobin/ENCODE2/Public/

This is NOT an official ENCODE release. For all the metadata, please refer to UCSC ENCODE portal:

http://genome.ucsc.edu/ENCODE/

To reduce file sizes, the quality scores were not recorded, and the read names were replaced with numbers.

The files are directly compatible with Cufflinks.
CSHL data is stranded (dUTP protocol) and Cufflinks has to be run with --library-type fr-firststrand
Caltech and HAIB data are unstranded and can be run with default --library-type.

STAR version: STAR_2.3.1u (2013/11/24)
Genome: hg19 + phiX + NIST ERCC spike-ins
Annotations: Gencode18

Please let me know if you have any issues or questions
Cheers
Alex

[Liy]

Thanks Alex!

Is these data normalized in some way?

L

[alexdobin]

Hi Liy,

at the moment I have posted only the alignments - BAM files, so there is no normalization of any kind. I was contemplating also making the signal (wiggle) tracks - these can be made in many different ways (normalization, unique- vs multi-mappers, etc.).

Cheers
Alex

[Maike]

mouse long RNASeq

Dear All,

there have been multiple questions about the ENCODE RNA-seq alignments on the UCSC portal. These alignments had been generated by a 3-year old version of STAR and use some non-conventional formatting (e.g. they are not compatible with Cufflinks).

To bring this data up-to-date, I have remapped it using the latest version of STAR. The new alignments use conventional formatting and should be compatible with most downstream software. Importantly, annotations are used to improve the mapping accuracy. The BAMs for all of the ENCODE phase 2 (2008-2012) long RNA-seq data can be downloaded here:ftp://ftp2.cshl.edu/gingeraslab/trac...t/ENCODE2/BAM/

This is NOT an official ENCODE release. For all the metadata, please refer to UCSC ENCODE portal:

http://genome.ucsc.edu/ENCODE/

To reduce file sizes, the quality scores were not recorded, and the read names were replaced with numbers.

The files are directly compatible with Cufflinks.
CSHL data is stranded (dUTP protocol) and Cufflinks has to be run with --library-type fr-firststrand
Caltech and HAIB data are unstranded and can be run with default --library-type.

STAR version: STAR_2.3.1u (2013/11/24)
Genome: hg19 + phiX + NIST ERCC spike-ins
Annotations: Gencode18

Please let me know if you have any issues or questions
Cheers
Alex

Alex, does anything like that exist also for the mouse RNASeq dataset?

I`m trying to get the counts per gene using HTseq on files I generated from the encode (CSHL long RNASeq) bam files (sorted & turned into sam using samtools). However only 13% of the reads actually map to features, regardless of the GTF fie I use. Could the reason be the same?

I appreciate any hints!

Maike

[alexdobin]

Hi Maike,

it's very likely that HTseq has troubles with the old BAM format. The main problem is that in this old format the mates were assigned the same strand (for better viewability on UCSC browser), however, this is not a standard convention for Illumina reads.

I am remapping the ENCODE CSHL mouse data to mm10 and Gencode M2 annotations (just released!), and will post the BAMs early next week.

Cheers
Alex

[Maike]

Thank you Alex, for answering and doing the work!
Maike

[alexdobin]

The re-mapped ENCODE2 mouse CSHL data is posted here:
ftp://ftp2.cshl.edu/gingeraslab/trac...AM/Mouse_CSHL/

[Maike]

mouse encode rnaseq

This is really helpful, thank you!

[Auction]

The re-mapped ENCODE2 mouse CSHL data is posted here:
ftp://ftp2.cshl.edu/gingeraslab/trac...AM/Mouse_CSHL/

Dobin, regarding the mouse-remapping, are you using this reference ftp://ftp2.cshl.edu/gingeraslab/trac...GencodeM2.tgz?

Thanks.

[alexdobin]

Dobin, regarding the mouse-remapping, are you using this reference ftp://ftp2.cshl.edu/gingeraslab/trac...GencodeM2.tgz?

Thanks.

Yes, this is correct.

[Auction]

Dobin

The reference in ftp://ftp2.cshl.edu/gingeraslab/trac..._GencodeM2.tgz only provides the GTF file and STAR indexed reference. Where can we download the fasta files for both mm10 and ERCC markers?

Thanks.

[rzhang]

Hi Alex,

I could not connect to the cshl ftp address and the link is broken. Could you please tell me where I can download the data now?

Many thanks,
Rui

Dear All,

there have been multiple questions about the ENCODE RNA-seq alignments on the UCSC portal. These alignments had been generated by a 3-year old version of STAR and use some non-conventional formatting (e.g. they are not compatible with Cufflinks).

To bring this data up-to-date, I have remapped it using the latest version of STAR. The new alignments use conventional formatting and should be compatible with most downstream software. Importantly, annotations are used to improve the mapping accuracy. The BAMs for all of the ENCODE phase 2 (2008-2012) long RNA-seq data can be downloaded here:ftp://ftp2.cshl.edu/gingeraslab/trac...t/ENCODE2/BAM/

This is NOT an official ENCODE release. For all the metadata, please refer to UCSC ENCODE portal:

http://genome.ucsc.edu/ENCODE/

To reduce file sizes, the quality scores were not recorded, and the read names were replaced with numbers.

The files are directly compatible with Cufflinks.
CSHL data is stranded (dUTP protocol) and Cufflinks has to be run with --library-type fr-firststrand
Caltech and HAIB data are unstranded and can be run with default --library-type.

STAR version: STAR_2.3.1u (2013/11/24)
Genome: hg19 + phiX + NIST ERCC spike-ins
Annotations: Gencode18

Please let me know if you have any issues or questions
Cheers
Alex

[alexdobin]

Hi Alex,

I could not connect to the cshl ftp address and the link is broken. Could you please tell me where I can download the data now?

Many thanks,
Rui

Hi Rui,

this is the new location of the ENCODE2 RNA-seq BAMs:

http://it-collab01.cshl.edu/shares/gingeraslab/www-data/dobin/ENCODE2/Public/

Cheers
Alex

[apredeus]

Hi Rui,

this is the new location of the ENCODE2 RNA-seq BAMs:

http://it-collab01.cshl.edu/shares/gingeraslab/www-data/dobin/ENCODE2/Public/

Cheers
Alex

this is so very useful, thank you very much, for mouse ENCODE CSHL data in particular! I was aligning these data: http://www.ncbi.nlm.nih.gov/geo/quer...i?acc=GSE39524, but it's been pretty painful since it's ABI SOLID platform...

BTW, Alex: have the mouse ENCODE data (CSHL long RNA-seq, the ones you have shared) been published yet?

[alexdobin]

this is so very useful, thank you very much, for mouse ENCODE CSHL data in particular! I was aligning these data: http://www.ncbi.nlm.nih.gov/geo/quer...i?acc=GSE39524, but it's been pretty painful since it's ABI SOLID platform...

BTW, Alex: have the mouse ENCODE data (CSHL long RNA-seq, the ones you have shared) been published yet?

Hi @apredeus,
our mouse paper is under review, however, these mouse data were released by ENCODE in 2013 and are now free of any restrictions, you can check this in the last column of this table:
https://genome.ucsc.edu/ENCODE/dataSummaryMouse.html