Question: Analysis of RNA-seq data of Ion Torrent
gravatar for Wang
14 months ago by
Wang10 wrote:

Hi, everyone. I'm a newbie to sequence data analysis. Yesterday I came across a sequence data created by Ion Torrent form GEO(GSEGSE89972).I plan to analysis the SRA data downloaded from GEO, but I just wonder how can I analysis these data properly. The same as those from Illumina?Are the pipelines or softwares are different from those used to processing Illumina RNA-seq data? Convert SRA files to fastq files by fastq-dump, QC by FastQC, Mapping by STAR,
... Or I should choose some other softwares? Thank you!

rna-seq • 545 views
ADD COMMENTlink modified 14 months ago by lshepard400 • written 14 months ago by Wang10

What do you want to do with the data, I mean what kind of analyses are you planning with it?

ADD REPLYlink written 14 months ago by lieven.sterck8.0k

I'd like do differential expression analysis.Or further pathway clustering,network analysis if possible.

ADD REPLYlink written 14 months ago by Wang10
gravatar for lshepard
14 months ago by
United States
lshepard400 wrote:

Most of the analysis is similar (e.g.: QC, count generation, differential expression etc...). Due to differences between the interments, you will typically see reads with lower Phred scores from the QC, but that's expected.

The primary difference is in mapping due to homopolymer reads. Check out this post for some discussion about this and suggestions on parameters. Additionally, you may consider using the "Two-Step Alignment" approach briefly discussed here for Ion Torrent data, but I suggest you search the web further for additional details on this method. Admittedly, the information about this is a bit less than the Illumina systems, but it's there if you dig in. In short this method suggests to use a spice-aware aligner to map the reads and then take the unmapped reads and use an aligner like Bowtie2 to complement the mapping (the poor mapping from the splice-aware aligner is usually due to the homopolymer reads). And then of course you would need to combine the BAM files and sort with samtools.

I would suggest to explore this further and benchmark some strategies and see what works best. There may be better approaches to deal with Ion Torrent data now (and perhaps quasi-mapping could do a good job? I have not tested this yet).

ADD COMMENTlink written 14 months ago by lshepard400

Thank you so much for those useful information! I really appreciate it!

ADD REPLYlink written 14 months ago by Wang10
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1354 users visited in the last hour