SEQanswers

Go Back   SEQanswers > Applications Forums > RNA Sequencing



Similar Threads
Thread Thread Starter Forum Replies Last Post
Spike-in for identification of samples hbn Illumina/Solexa 4 08-24-2011 07:40 AM
spike-in data set nclement Bioinformatics 5 12-02-2010 07:39 AM
Spike sequence for ERANGE pibbo RNA Sequencing 0 06-03-2010 06:04 PM
Spike sequence for ERANGE pibbo Bioinformatics 0 06-03-2010 11:42 AM
(a) Background Read Distribution (b) Spike-ins ANJAN PURKAYASTHA Bioinformatics 0 03-05-2009 07:02 AM

Reply
 
Thread Tools
Old 08-28-2011, 11:10 PM   #1
Suthira
Junior Member
 
Location: Muenster, Germany

Join Date: Jul 2010
Posts: 1
Default ERCC Spike-In Control

Hi everyone

we're hoping to start an RNASeq run on the Illumina HiSeq SQ in the next few weeks. I recently came across the ERCC Spike-In Control, which lets you add known amounts of mRNA to your sample.. I'm interested in using the ExFold mixes, which are two mixes with different amounts of the same mRNAs.

It looks like it could be really useful! I've had a look at the "Synthetic spike-in standards for RNA-seq experiments" paper, but I have little background in bioinformatics... what they've done seems too complicated for me to repeat.. Does any one know a program which can take advantage of spike-in controls (eg. for normalisation, checking quality of the run)?

Any advice will be greatly appreciated

Thanks in advance,
Suthira
Suthira is offline   Reply With Quote
Old 06-21-2012, 11:31 AM   #2
Frazzled
Junior Member
 
Location: LA

Join Date: Jun 2012
Posts: 2
Default

Hi,

I am going to start an RNASeq run on the Illumina HiSeq SQ with 50 bases paired-end sequencing in the next few weeks. And I also came across the ERCC ExFold Spike-In Control. I was very excited to see your post online, could you please share with me your experience with the ERCC control? Is there any caveats when combining them into RNA sample in making the cDNA library and how well this control performs? Any advise will be highly appreciated. Thank you in advance!

Frazzled
Frazzled is offline   Reply With Quote
Old 05-13-2013, 03:54 PM   #3
danwiththeplan
Member
 
Location: Auckland

Join Date: Sep 2011
Posts: 72
Default

Ion torrent has a plugin that automatically analyses ERCC controls. No idea how well it works. I have also written some scripts to analyse ERCC controls from illumina data, and so has this guy:

https://github.com/metalhelix/spikein_analysis

Again no idea how well it works.

I've written some scripts (in python and R) to do some really basic analysis and produce a graph of observed vs expected (Ie log2 FPKM values vs known input concentrations). There's a lot more you could do with these controls (e.g. normalisation, looking at antisense vs sense transcripts to see how "strand specific" strand specific data really is).
danwiththeplan is offline   Reply With Quote
Old 06-04-2013, 08:32 PM   #4
Macrophage
Junior Member
 
Location: Quebec City

Join Date: Jun 2013
Posts: 4
Question idem

Hi Frazzled! This is probably too late but I would like the discussion to be open again. 1 year after you, I have exactly the same question and have some troubles to find someone who has already used it for RNAseq. Everybody around me tells, this is absolutely needed to compare results from different platforms but nobody use it or at least knows someone that would have been eared about a guy who would have already used it.
This approach is strongly recommended by the ENCODE consortium (http://genome.ucsc.edu/ENCODE/protoc...dards_V1.0.pdf + https://www.google.ca/url?sa=t&rct=j...47380653,d.aWM + http://nxseq.bitesizebio.com/article...-its-solution/).

However, I could only find 3 papers seriously considering this and performing some pilot experiments (http://genome.cshlp.org/content/21/9/1543.long + http://www.sciencedirect.com/science...92867412012263 + http://link.springer.com/article/10....427-013-4437-9)

I contacted both Illumina and Epicentre and they don't really care about this. Epicentre just told me that one customer lost half of his reads in spike-in sequencing (I guess he/she should have put a huge amount, I have no information on it).
"[...] we do not normally recommend them, and I just spoke with a customer overseas who did use them and it completely "swamped out" the data from the sample such that the spiked RNA provided the bulk of the data. Remember that ScriptSeq uses a very minimal amount of RNA (500 pg to 50 ng) and we'd be worried about how much of the data generated would be from the bona fide sample versus the spiked in RNA."

Staff Technical Applications Scientist
Epicentre (an Illumina company)
(800) 284-8474, ext 66119
608.442.6119
Website: www.epicentre.com

My conclusion is that one who wants to use it should perform a pilot experiment (what should actually be done for any RNAseq analysis). I know because of time and money, this is not always possible. In my case, I have neither time nor money for pilot experiments so I'll go (with deep regrets) to library construction without spike-in controls and assess like many others that the coverage and global expression levels are not significantly variable between my samples (what I guess is not true).
Anyway, there is no perfect experiment and there are probably many bad things we are doing and we don't know yet. For sure, RNAseq standards are not fixed forever.

As I'm very new in RNAseq, I would be very glad to ear different opinions if somebody already included these controls.

Last edited by Macrophage; 06-04-2013 at 10:06 PM.
Macrophage is offline   Reply With Quote
Old 06-04-2013, 08:45 PM   #5
Jon_Keats
Senior Member
 
Location: Phoenix, AZ

Join Date: Mar 2010
Posts: 279
Default

We include the ERCC in most experiments now. We see 0.2-1.5% of all reads coming from the spike-ins. Personally I see a significant amount of value in these reagents and I highly recommend them to collaborators now as we design new projects.
Jon_Keats is offline   Reply With Quote
Old 06-04-2013, 10:00 PM   #6
Macrophage
Junior Member
 
Location: Quebec City

Join Date: Jun 2013
Posts: 4
Question Glad to ear this!

Hi Jon! Thank you very much for your post! I would be interested in the amount of ERCC you add to your samples. Do you directly use the recommendations from Ambion or do you have adapted protocols depending on the sample type. In my case, this will be human macrophages. Libraries will be prepared with Ribozero + ScriptSeqV2 (no polyA selection) from 1g total RNA.
If you have any advice, I need arguments to convince my PI this would be a good idea to consider this. I would really like to include them but we're not very large about the number of reads we can get from each sample (money limitations) so we're afraid about adding too much ERCC in our preps.
Macrophage is offline   Reply With Quote
Old 06-06-2013, 08:21 AM   #7
NextGenSeq
Senior Member
 
Location: USA

Join Date: Apr 2009
Posts: 482
Default

It's too bad the ERCC Spike-In Controls cost so much. $2100 for both control mixes is really pretty expensive.
NextGenSeq is offline   Reply With Quote
Old 06-06-2013, 09:58 AM   #8
Macrophage
Junior Member
 
Location: Quebec City

Join Date: Jun 2013
Posts: 4
Smile

A bit expensive but not so bad but for such amount of pure synthetic RNA. Also you can use only one mix for 1130$ (http://products.invitrogen.com/ivgn/product/4456740).

Since there is enough for at least 5 flow-cell experiments (this count is with 1g total RNA samples if you put 50 samples max / flow cell), you can buy one kit for several collaborating labs.

Personally, I find it pretty cheap compared to cBot and SDS Illumina reagents...
Macrophage is offline   Reply With Quote
Old 06-06-2013, 01:10 PM   #9
danwiththeplan
Member
 
Location: Auckland

Join Date: Sep 2011
Posts: 72
Default

Yes I've talked to the reps complaining about the cost, and suggesting they sell them in smaller pots, they seemed to be receptive.
danwiththeplan is offline   Reply With Quote
Old 06-06-2013, 01:55 PM   #10
jparsons
Member
 
Location: SF Bay Area

Join Date: Feb 2012
Posts: 62
Default

I think that there is little reason to follow the protocol with regards to how much spike-in control you add. Since the pools are set up as a dynamic range from 1-1million, the top 3 spike-ins are often more heavily represented than the most abundant genes in your sample. If you add 1/8 as much spike per sample, you can still easily determine your lower limit of detection and fold change response, you just lose the ability to detect a few of the spike-ins at the lowest range.

It's advisable to tailor your use of the spike-ins to your experimental needs. For example, the short poly-A tails of the ERCCs mean that only a fraction of your controls will pass a poly-A selection. If you are not planning to use the ERCCs to compare efficiencies of various enrichment procedures or looking to identify the mRNA:totalRNA fraction between multiple samples , you should not spike in the ERCCs before poly-A selection. Ribodepletion doesn't affect ERCC counts nearly as much.

**I have only done analysis on the NIST SRM ERCC controls, not the Ambion mixes. I'm therefore not exactly familiar with their suggested protocols beyond a brief perusal.
jparsons is offline   Reply With Quote
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 01:42 PM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2019, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO