I have 100 RN A seq samples and i have mapped them using tophat. but the alignmnet is so so.
like given below:
Left reads:
Input : 24818
Mapped : 1562 ( 6.3% of input)
of these: 122 ( 7.8%) have multiple alignments (20 have >20)
Right reads:
Input : 24818
Mapped : 1851 ( 7.5% of input)
of these: 130 ( 7.0%) have multiple alignments (20 have >20)
6.9% overall read mapping rate.
Aligned pairs: 1062
of these: 79 ( 7.4%) have multiple alignments
80 ( 7.5%) are discordant alignments
4.0% concordant pair alignment rate.
how much threshold i should set to select the rna sequnce samples.Should i keep it 1 million or 2 million .If i set it to be 2 million my 32 sequences are deleted.but if i select it to be 1 million almost 20 are deleted.
I want to know whether i can keep threshold =1million as i dont know how many reads are considerd to be good alignmnet reads
in transcriptome dont try to enlarge your reads count but try to filter out common sequence.. focus on genes and expressions
First, it would be good to know a bit more about your experiment: which specie do you study, what kind of RNA, which genome did you use, how where the libraries generated and sequenced... Then if it's standard mRNA sequencing in a common vertebrate specie, your total number of reads is quite low and would give a bad covering of the transcriptome (24000 reads for 25000-30000 genes i.e. less than one read per gene). I woud suggest to less multiplex your samples, filter the fastq files (with cutadapt for example) for repeated sequences after performing a quality assessment with fastQC and then redo your alignments with something like 20-30 million input reads per sample.
With your number of reads, I guess that it could be microbial RNA-seq, but the same quality checks apply, you may just be able to obtain nice results with a lower number of input reads (scaled down to the number of genes and the genome size of your specie of interest). What you should aim at is to have the highest mapping percentage possible, and maybe to filter out reads with low mapping quality. Good luck!
Hi,
1. The number of reads sounds super fishi to me. ARe you sure that you downloaded the full data sets?
2. The mapping rate sounds also very low. Are you sure the mapping actually worked? Usually I would expect something around 80%.
I agree with both Frabrice and Fritz but your reads and mapping rate is terribly low, I see 80-90% in Drosophila and shoot for about 60m reads per sample...More info on your organism and design would be helpful..
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA