Process RNA-Seq Data with specific sequences / not whole reference genome
I have specific sequences that I should quantify in RNA-Seq data. First, I blast the sequences against the reference genome. Now these data have to be compared with the reads of the RNA-Seq analyses. I used the specific sequences in Bowtie 2 as reference genome (fa format) and mapped the RNA-Seq data with them. It worked, but is this the right way? I also have to count the mapped reads, but it doesn't work with htseq-count because the identifier is not unique. How can I solve this problem?
• 155 views
•
link
written
by
Heinrich_M •
0 answers
No answers yet.
Log in to answer this question.
More posts like this
-
Htseq-count output file having a high number of __not_aligned
written by Nemo •
-
HTSeq-count: proper pairs in bam file
written by mario_de_manner •
-
small RNA seq data from human cell line only give me 0 counts
written by leandroteixeira •
-
Comparison of transcripts of different parasite stages - graphic output
written by Heinrich_M •
-
Counting single end reads
written by valopes •
-
Low counts with htseq-count?
written by longoka •
-
Anyone knowns piRNA sequencing analysis strategy?
written by Pinji Lei •
-
Anyone knowns piRNA sequencing analysis strategy?
written by Pinji Lei •
-
counting reads over repeats
written by R •
-
Should Rna-Seq Reads Map To One Strand Of Cdna Reference?
written by chjiao3456 •
Why do you use BLAST rather than Bowtie? You can use it of course just remember that BLAST is a local aligner. I didn't understand how you tried to run htseq-count, I have a feeling it expects sam file and you gave it something else
I did a blast alignment and then converted it to a bed file by adding values. With Bed-to-GFF I had the GFF for htseq. As Bed file I had the reads mapped against the reference genome.
but, as you can see, i don't know how to use it well yet and maybe that's not a good approach.
Mapping to a reduced reference is always an invitation for getting alignments that may be non-specific. Aligners will try to align at their best and with a reduced reference that can cause issues.
Always align against the reference genome and count from there on.
Ok thank you very much for this comment. Than I will start to map the specific sequences against the reference genome with Bowtie and than I have to count them. But how can I count them. I'm confused how to manage the genome, the sequences and the reads.
It's pretty straightforward, map the fastq files to the genome, run htseq-count and analyze only the features (genes?) of interest.
Then I have to create a GFF file from the sequences (I already did that, but then an error message is displayed Fatal error: Unknown error occured [bam_sort_core] merging from 1 files and 1 in-memory blocks... Error occured when processing GFF file (line 3 of file /home/galaxy/galaxy/database/files/051/dataset_51952.dat): Feature W03_711bp does not contain a '-i' attribute [Exception type: ValueError, raised in count.py:77] )
I guess it's because the BAM file contains gene_ids, but the gff does not. And I'm dealing with repetitive elements. Is there an option to count the sequences without greating a gff file?
BAM file shouldn't contain gene ids just chromosomes and positions. The genes (or features) are defined in the gff file
Okay, my fault. I guess the BAM file is good. But the gff is not. Now these files are not comparable. Is there an option to count the sequences without greating a gff file?
Can you get a GTF/GFF from the source where you got your reference from? Things should match then.
Yes I have the GFF File, but I guess it just contains the genes and I'm searching for repetitive elements. And I have still the error with the gene_id when matching the Genoic GFF against the BAM file. Is there an option to count the sequences without greating a GFF file?