small RNA sequencing

[sameet]

Hi,

We recently obtained Illumina small-RNA RNA-seq data for S. pombe. I believe that the next step should be use of BowTie to map the reads back to the reference genome. Is there any specific parameters that I need to adjust because of this being specifically small-RNA RNA-seq?

[NicoBxl]

what is the purpose of your research ? mirna analysis ? de novo assembly ?

[sameet]

Yes, our primary aim is to study changes in micro-RNA profiles across a spectrum of genetic backgrounds.

[NicoBxl]

ok,

Which sequencer ? Illumina GA II, 454, ... ?

[sameet]

Illumina GA-II. And because this is small RNA during library preparation we used the RNA between 17 - 40 bases.

[NicoBxl]

the first step is the adapter trimming. After that you can align the reads to the genome.

[sameet]

Can you give me the names of programs and possible parameters for the adaptor trimming. I plan to use BowTie for the mapping of reads to the genome. I am only worried about one thing. As i am looking for small-RNAs, i do expect a lot of repetitive regions, how do i uniquely map them?

[NicoBxl]

for the adapter trimming, you've to aligne the adapter sequence to the 3' part of the reads . Novoalign do that I think.

For bowtie, a parameter ( I don't remember which on ) exists to specify the max number of matching position .

[mmartin]

You may want to try to use my tool cutadapt for the adapter trimming.

[sameet]

@mmartin, I downloaded the code for cutadapt, however, it is not very clear how to use it with Illumina. Can you point me in correct direction?

[mmartin]

If you have a FASTQ file, run:

cutadapt -e ERROR-RATE -a ADAPTER-SEQUENCE input.fastq > output.fastq

ERROR-RATE is a real number between 0 and 1 such as 0.05.
I will update the documentation soon. There is no QSEQ support, yet.

[sparks]

Novoalign for miRNA

Hi Sameet,
Novoalign will trim adapters and align miRNA in one go. It's also pretty fast even free version with no multi-threading as reads are usually aligned with at most one mismatch.
Novoalign also looks for precursor location, giving a score and location for nearby complementary alignment.
Steps:
1. index genome
novoindex ref.idx ref.fasta
2. Align the miRNA
novoalign -d ref.idx -a -m -t 30 -f reads.fastq

Colin

for the adapter trimming, you've to aligne the adapter sequence to the 3' part of the reads . Novoalign do that I think.

For bowtie, a parameter ( I don't remember which on ) exists to specify the max number of matching position .

[sameet]

Hi Sparks,

Thanks that was useful. I will look into this and keep you posted.

[sameet]

Hi Sameet,
Novoalign will trim adapters and align miRNA in one go. It's also pretty fast even free version with no multi-threading as reads are usually aligned with at most one mismatch.
Novoalign also looks for precursor location, giving a score and location for nearby complementary alignment.
Steps:
1. index genome
novoindex ref.idx ref.fasta
2. Align the miRNA
novoalign -d ref.idx -a -m -t 30 -f reads.fastq

Colin

Hi,

I was wondering about one more question. If the adaptors are trimmed, would you get the final 'usable' (for the lack of better word) sequences of different lengths?

[mgogol]

Yes.

You can also use fastx_clipper to trim and fastx_collapser to collapse the reads if you want.

There's a few web based tools you can plug in your collapsed sequences and counts and get some results... I'm trying them now and reading the papers just to see how they are and what they do, I might do something myself also, but it's easy enough to plug and chug. mirAnalyzer, mirTools, and DSAP.