search

kingsfordgroup / sailfish

Rapid Mapping-based Isoform Quantification from RNA-Seq Reads
http://www.cs.cmu.edu/~ckingsf/software/sailfish
GNU General Public License v3.0
124 stars 45 forks source link

Sailfish-Paired end single replicate study #90

Closed misrak closed 8 years ago

misrak commented 8 years ago

Hello All,

I wanted to know if I can use Sailfish for single replicate study for three different conditions.

Thank you!

rob-p commented 8 years ago

Hi @misrak,

I presume your goal is to test for differential gene / transcript expression across conditions? Sailfish will give you accurate estimates of transcript expression in each sample (i.e. in each single replicate of each condition). Additionally, you can use the --numBootstraps option to generate bootstraps for each sample which will allow for an estimate of the technical variance of the abundance estimates. However, I'll parrot the now-standard warning — you should be very wary of attempting to make DE calls with only a single replicate per-condition. This advice is independent of how accurate your quantification tool is (i.e. this advice holds even if you perfectly estimate transcript abundance in each sample), since a single replicate per-condition prevents you from estimating each condition's biological variability. In this case, downstream DE tools (like DESeq2) may treat different samples as replicates for the purpose of estimating variability (I refer you to the excellent DESeq2 vignette for more details). TLDR — Sailfish will quantify these samples just fine, and you can run DE tools on the results, but best practices highly-recommend having multiple replicates per-condition for the purposes of DE.

misrak commented 8 years ago

Thank you so much. :)

Sent from Outlook Mobilehttps://aka.ms/blhgte

On Mon, Mar 21, 2016 at 6:50 AM -0700, "Rob Patro" notifications@github.com<mailto:notifications@github.com> wrote:

Hi @misrakhttps://github.com/misrak,

I presume your goal is to test for differential gene / transcript expression across conditions? Sailfish will give you accurate estimates of transcript expression in each sample (i.e. in each single replicate of each condition). Additionally, you can use the --numBootstraps option to generate bootstraps for each sample which will allow for an estimate of the technical variance of the abundance estimates. However, I'll parrot the now-standard warning - you should be very wary of attempting to make DE calls with only a single replicate per-condition. This advice is independent of how accurate your quantification tool is (i.e. this advice holds even if you perfectly estimate transcript abundance in each sample), since a single replicate per-condition prevents you from estimating each condition's biological variability. In this case, downstream DE tools (like DESeq2) may treat different samples as replicates for the purpose of estimating variability (I refer you to the excellent DESeq2 vignettehttps://bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.pdf for more details). TLDR - Sailfish will quantify these samples just fine, and you can run DE tools on the results, but best practices highly-recommend having multiple replicates per-condition for the purposes of DE.

You are receiving this because you were mentioned. Reply to this email directly or view it on GitHubhttps://github.com/kingsfordgroup/sailfish/issues/90#issuecomment-199289785