a question in using Flux Simulator

[alittleboy]

I am using Flux Simulator to simulate RNA-Seq experiment, and after running

flux-simulator -p /HS_simu1.par

I got the following error message:

[ERROR] Error while preparing sequences: Problems reading sequence GL000191.1: pos -1, len 171, check whether chromosomal sequence exists / has the correct size

Can I know which file I should check for chromosomal sequence? Thanks!

[alexdobin]

I have just had a similar problem with Flux Simulator - I think this problem is caused by a transcript (in your ENSEMBL(?) GTF) that belongs to GL000191.1 "non-chromosomal" contig, while you only provided Flux with "chromosomal" fasta files. The easiest way to solve it is to filter all the "non-chromosomal" annotations from the GTF, or you can add the "non-chromosomal" fasta files.

[alittleboy]

I have just had a similar problem with Flux Simulator - I think this problem is caused by a transcript (in your ENSEMBL(?) GTF) that belongs to GL000191.1 "non-chromosomal" contig, while you only provided Flux with "chromosomal" fasta files. The easiest way to solve it is to filter all the "non-chromosomal" annotations from the GTF, or you can add the "non-chromosomal" fasta files.

Thank you alexdobin for the suggestions! Can I know how you solved this problem (i.e. modifications of the GTF file...)?

[alexdobin]

Since I did not need annotations on non-chromosomal scaffolds, I simply removed all the non-chromosomal entries from the GTF file. For ENSEMBL GTF, the chromosomes are 1-22,X,Y,MT, so you could simply do:
grep ^[0-9XYM] ENSEMBL.gtf > ENSEMBL.chrOnly.gtf

On the other hand, if you want to simulate RNA-seq from non-chromosomal scaffolds, you will need to downolad the "nonchromosomal" fasta file from ENSEMBL (e.g. ftp://ftp.ensembl.org/pub/release-72...omosomal.fa.gz), split the file into separate fasta files for each of the scaffolds, and add these fasta files to the GEN_DIR directory.

[CompBio]

I've run into a similar problem with whole chromosomes. At the time I assumed the simulator may try to grab sequence past the end of a chromosome when there are genes near the end of a chromosome. My simple workaround was to pad the end of each chromosome sequence with enough N's to accommodate my desired read size.

It's possible the same solution would work with contigs/scaffolds if you don't want to eliminate them from your simulations.