[jp.]

Hi
may I get your views on this thread of my question:

http://seqanswers.com/forums/showthread.php?p=112098#post112098

That is strange. When you run cufflinks, are you running it with -g (use reference transcripts and assemble new ones) or -G (only use reference transcripts)? I am no expert but I imagine this error could arise depending on how you mapped to the known transcripts.

[jp.]

Hi all
I am having the same problem...
Warning: couldn't find fasta record for 'NT_166469'! This contig will not be bias corrected.
Warning: couldn't find fasta record for 'X'!Warning: couldn't find fasta record for 'Y'! This contig will not be bias corrected.
Initially, I used ensemble ref and got this warning and then I used UCSC genome.fa. Both gives me the same error.
Why ??

Psikon,

Your hg19 reference file has chromosome names in the format 'chr1' so that is the format they would be in the BAM file(s) generated by TopHat. Whereas the Ensembl GTF you showed in the example has chromosome names as just the numbers (e.g. '11'). Since the alignments in your BAM file(s) are to references named 'chrN' and the annotations in the GTF file are to references named just 'N' Cufflinks naturally won't be able to match them. You need to ensure that the chromosome names in the reference FASTA you use to map exactly match the names in the annotation file. This is typically best done by getting your reference and annotation from the same source.