How to obtain full length RNA transcript sequence

[Annibal]

Hi everyone,
i'm new to this kind of tasks so, please be patient!
I'm trying to create a blast DB using the RNAseq data from ENCODE.
I've downloaded both the FASTQ reads and the .bam/bai files.
I need the fasta sequences of all the full length transcripts: is it possible to extract/obtain them from the BAM file?
Alternatively should i try to do a de novo assembly using Trinity?
Thanx a lot.
Regards,

Davide

[Annibal]

I thought this task would have been easy or at least possible since i have the reads aligned to the ref genome (homo sapiens)
Anyone can help?
Thanx

[Simon Anders]

Which BAM files are you talking about? ENCODE has many.

Why do you want to make your BLAST data base from RNA-Seq reads rather than simply from, say, the cDNA FASTA file from Ensembl?

[Annibal]

Which BAM files are you talking about? ENCODE has many.

Why do you want to make your BLAST data base from RNA-Seq reads rather than simply from, say, the cDNA FASTA file from Ensembl?

I'm talking about BAM file of the human total RNA extract from CSHL Long RNA seq.

I don't use Ensembl data because cDNA FASTA from Ensembl does not contain all the transcript (i guess) but only "known, novel and pseudogenes" as stated on their website

Moreover i will probably repeat this task using RNAseq data from cell in particular conditions

Thanx a lot.

Davide

[Simon Anders]

What you want to do is called reference-based (as opposed to: de-novo) transcript assembly. A tool commonly used for this purpose is cufflinks:

Roberts, Pimentel, Trapnell, and Pachter:
Identification of novel transcripts in annotated genomes using RNA-Seq
Bioinformatics (2011) 27 (17): 2325-2329.
doi:10.1093/bioinformatics/btr355

However, before doing this yourself, you may want to check whether the ENCODE people have not already done this analysis. It seems obvious that they would do this.

I still wonder what you would need a database of all transcripts for. Instead of blasting against it, you can always blast against the genome.

[Annibal]

What you want to do is called reference-based (as opposed to: de-novo) transcript assembly. A tool commonly used for this purpose is cufflinks:

Roberts, Pimentel, Trapnell, and Pachter:
Identification of novel transcripts in annotated genomes using RNA-Seq
Bioinformatics (2011) 27 (17): 2325-2329.
doi:10.1093/bioinformatics/btr355

However, before doing this yourself, you may want to check whether the ENCODE people have not already done this analysis. It seems obvious that they would do this.

I still wonder what you would need a database of all transcripts for. Instead of blasting against it, you can always blast against the genome.

Thank you.
I've taken a look at cufflinks, just the manual, but i did not find th FASTA file of the transcript as an output file of some task. Cufflinks instead talk about gtf file as an output (that does not contain the FASTA sequence of the transcript). I'll take a better look to the program.
I've also read just yesterday this interesting article:
"Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks" nature protocol

If i blast in the genome i lose informations that are in the RNA sequence and not in genome (ex. sequences in the transposable element that are not integrated in the genome...)

Thanx again.

[Simon Anders]

You use the GTF file to produce the cDNA FASTA file from the reference FASTA file. This is a simple exercise in script programming.

[Annibal]

I've taken a look at GTF specs.
Yes, it is.
Thanx

[oxydeepu]

Hi i saw the thread..
Can i get the logic for the program to create transcripts from genome file.
how it differ based on orientation. i mean reads which have positive and negative orientation.
Thank you.
Deepak

[swaraj]

Refer to my earlier post to get fasta from Cufflinks GTF.

http://seqanswers.com/forums/showthread.php?t=18369