Genomes FASTA/GTF files needed to run the evaluation about ultra HOT 4 CLOSED

Awesome, thank you for your prompt responses. I could simulate the three datasets.

from ultra.

Hi @FadelBerakdar

For human, I am using GENCODE v42 instead of GENCODE v34, and for drosophila release 108 instead of 97. Is this kind of upgrade in versions is not recommended?

I don't know what the updates are, but if its improved annotations, it should be fine. uLTRA is not dependent on any specific version.

In the experiment.json file, "TRANSCRIPTOME" : "/home/kris/source/ultra/data/all_transcripts_ENSEMBL.txt", is this the same as GENCODE transcripts.fa ?

No, those are ENSEMBL biomart transcripts. You can ignore those. The relevant annotation we used are in https://github.com/ksahlin/ultra/blob/master/evaluation/Snakefile#L219 (those specific annotation versions should be given in the paper; v34 and release 97)

For the SIRV, I am not sure which GTF file I should use,

Please use the SIRV_isoforms_Lot00141_multi-fasta-annotation_C_170612a.gtf. The 'C' stands for complete annotation. I and stands for incomplete and O stands for overannotated.

from ultra.

Hi @ksahlin

Thank you for your respsone.
The config['TRANSCRIPTOME'] is used by the snakefile in https://github.com/ksahlin/ultra/blob/master/evaluation/Snakefile#L108 and is passed to simulate_reads.py which I guess expects an argument sequence_material in fasta format? as in https://github.com/ksahlin/ultra/blob/master/evaluation/simulate_reads.py#L152

My assumption is that argment is a FASTA file contains the cDNA sequences that are relvant to the GENCODE human geneome ? kindly clearify

Thank you,
Fadel

from ultra.

Ah right! We used the ensemble biomart transcripts (cDNA) obtained from the biomart ensemble webpage to simulate FSM and NIC reads. (the simulated data in our benchmarks).

sequence_material in fasta format?

yes, the ensemble biomart transcripts (cDNA) in fasta format

My assumption is that argment is a FASTA file contains the cDNA sequences that are relvant to the GENCODE human geneome ?

The ensemble biomart transcripts are relevant to simulate reads.

from ultra.