[Dolphin22]

Thanks for your reply.
I did use the Helipshere software for the analyis so the artifacts should be removed automatically. But I will check the nucleotide distribution of my unaligned sequences allos manually.

[singlemoleculer]

The standard Helicos filtering software gets rid of these artifacts automatically. If you don't have access to the software, you can do manual filtering. Very rich in AT means A+T>90%. These are usually caused by problems in sample preparation. Any such sequences should be discarded unless you are working with a very AT-rich set of sequences like Plasmodium in which case you have to adjust the cutoff. For CTAG, you should look at the dimer frequencies. If CT+TA+AG+GC>70%, you should also discard. These sequences are usually caused when 2 DNAs are too close on the flow cell and the signals overlap. The filtering software actually has a more complex algorithm but this is good enough for most purposes.

[Dolphin22]

@ singlemoleculer
Thanks for your advices. I checked my sequences but I am not sure about some points:

What does very rich of A's and T's mean? Do you have a percentage? The same with the low number of G's.
Of course I find CTAG's but where is the limit?

I also find many sequences with long stretches of polyA's polyT's? Is this also an indicator for bad sequencing?

[Elcannibal]

This horse deader than dead when ILMN's lawyer team destroyed the community college kid who worked for free this week...

[singlemoleculer]

Dolphin22-
Your low aligned yield rate could be caused by many things but you can frequently diagnose some of the problems by looking at the sequence of what doesn't align. If they are very rich in either As or Ts, it was probably a sequencing problem. If short on Gs, it probably means the formaldehyde killed the Cs so you can't see the Gs. If an excess of CTAGs or portions thereof (standard Helicos base order of addition), you didn't filter properly. You may be able to sort out other issues by looking for common features of non-aligning sequences. For example, do they align to other species or does the same sequence keep appearing?
S

[Kletz_AUS]

Dolphin22 curious what lab you work in? Thank you.

[Dolphin22]

Hey everybody,

good to have some people already having some experience with Helicos.

In our Lab we are working on an RNA binding protein and try to figure out putative target RNAs. Therefore we did RIPs and subsequently sequenced them with Helicos.
But we have pretty low number of aligned reads (up to 21 %). We suppose this is due to the treatment of the samples with formaldehye. Doe anyone of you has experienced similar problems or has an alternative solution?

[Kletz_AUS]

Any updates from any of the users who were/are still working on the Heliscope?

[Lee Sam]

Gee, I asked Helicos some time back the same question and got the same answer. Helicos does work. It does a marvelous job on ChIP-Seq-RNA Seq and small genomes. The PacBio numbers are not that impressive right now. I look at them and place them about where Helicos was about three years ago. SMS is hard. It's definitely the future but keeping track of singular molecular events is a beotch. SMRTBell is a very clever way of getting around high error rates that SMS is prone to but it negates the best thing about SMS, you have to amplify. And you have to start with 20,000 times more DNA. Not a very good trade off but the technology will advance.

I'm really interested in seeing what PacBio can output for RNA-Seq since their sequencing protocol appears to output only like 90k strands (maximum) per chip (and that's assuming 100% well efficiency, which is for sure not the case). Last time they visited us we were told that they had something for doing RNA-Seq so only time will tell.

[polytoo]

I asked a PacBio rep why he thought their SMS will succeed when Helicos failed. He said because their instrument will work.

Gee, I asked Helicos some time back the same question and got the same answer. Helicos does work. It does a marvelous job on ChIP-Seq-RNA Seq and small genomes. The PacBio numbers are not that impressive right now. I look at them and place them about where Helicos was about three years ago. SMS is hard. It's definitely the future but keeping track of singular molecular events is a beotch. SMRTBell is a very clever way of getting around high error rates that SMS is prone to but it negates the best thing about SMS, you have to amplify. And you have to start with 20,000 times more DNA. Not a very good trade off but the technology will advance.

This may be history repeating itself. Initially people were excited about Dupont's 1st gen sequencer (which didn't work) and then ABI pulled it off successfully.

Dupont's Genesis 2000 did work and worked quite well. Where do you think ABI got dye terminators? But you are right, it might be history repeating itself in spades. Superior technology sometimes does not win.

[NextGenSeq]

I asked a PacBio rep why he thought their SMS will succeed when Helicos failed. He said because their instrument will work.

This may be history repeating itself. Initially people were excited about Dupont's 1st gen sequencer (which didn't work) and then ABI pulled it off successfully.

[bioits]

I think Paul knows NASDAQ rules. I guess what he really meant is he doesn't know why Helicos' marketing team didn't do a good job. After all, the technology itself is good enough for some applications.

They got delisted because their stock price didn't have a price level over $1 per NASDAQ rules.

I have experience with Helicos mRNA-Seq (having just submitted a paper using the technology).

[Lee Sam]

As of this week we have completed over 1,500 ChIP-Seq samples on the Helicos. A few nanograms, no library creation, just a little poly A tailing and zoom, 9-14 million reads on 46 samples every 14 days (we like to relax before we fill it up again.).

Just don't ask me why they just got delisted on NASDAQ and are trading at 22 cents. I have no idea. ;-( Keith Robison spells it out pretty well in the other thread. Their marketing should have been something like "It's all about the message" and focused on ChIP-Seq, RNA-Seq, DGE and all of the above from paraffin, nanogram amounts, no amplification, no bias.

They should have never chased after the human genome. Wrong tool for the job.

-Paul Morrison

They got delisted because their stock price didn't have a price level over $1 per NASDAQ rules.

I have experience with Helicos mRNA-Seq (having just submitted a paper using the technology).

[polytoo]

Paul,

do you know if there is any published or publicly available ChIP-seq dataset from independent groups?

Since May of 2009 we have worked on 26 different projects in 16 different labs. Virtually all are "any day now" in the publishing pipeline. Until then I can't talk about them or even name the labs.

If a request came into my lab specifically for ChIP-Seq results I could pass the request on to labs that are close to publishing. One also might be able to get a handle on the right folks if you did a search on "new epigenetics center" and my Institute.

[Chipper]

Paul,

do you know if there is any published or publicly available ChIP-seq dataset from independent groups?