I'm considering doing RNASeq for the first time on patient samples, and a company called Novogene is offering a really cheap price for RNASeq (1/3 of the cost at my institution's Core). Does anyone have experience with this company? Why is their price so cheap? Should I be concerned about data quality? I would appreciate any comments.
We use Novogene for all our sequencing, and have never had any problem with them - we've done all sorts - low pass whole genome, RNA-seq, NET-seq, smRNA-seq, iCLIP - granted, they only made the libraries for the standard RNA-seq and smRNA-seq, and other times, they've sequenced libraries we've made.
As for comms/customer service side - when we first started using them, we had a very good rep that covered our area and when ever we needed anything dealing with , we just when through her. Then she moved on, and for a while things weren't so good, we'd go ages without hearing from them, stuff would take a long time and communications seemed difficult. Now we have a good dedicated rep again. Nice guy. Always help. So I think if very much depends on who your rep is an building a personal relationship with them.
As for why it's so cheap - it's been pointed out - they have a lot of very high-throughput sequencers. Your sample will be one of many on a given lane. I also have a suspicion that they handle such a high throughput, that it would make sense for them manufacture their own library kits. You'll note that nowhere on the website does it say whose kits they use. If you ask them to break it down the RNAseq cost into sequencing and library, you'll see that the library prep costs really are very low.
We had a very bad experience with Novogene which made us realized that they are purposed to make money instead of try to help us generate good data.
In the primary library construction, 4 of 6 from our samples failed due to high rRNA content. Then they did a reconstruction and sequencing without letting us know that 4 samples still have >40% of rRNA. Then the results seems like messed up. Extremely high Ribosome related genes were mapped and enriched.
We were arguing with them for months and what they said is that nothing is wrong with the library construction. They asked us if something in the sample affected the mRNA pull-down (Our QC results showed that RNA quality is good).
Then we asked for a re-construction, they told us 4 samples have high rRNA again in this time again.
Then they told us that what they did is that the rRNA is bioinformatically removed after sequencing and they don't know if they are high after library, LOL.
They kept saying that our sample have problem and asked us to pay the money, at the same time, they refused to release the data.
We would never do any more with them and neither our university.
From the data I've seen from them the quality is usually OK to good, so no worries on that side.
A downside is that communication with them can sometimes be difficult. And (but counts probably for all service providers) is that as long as the analysis/sequencing you're requesting is main stream things will go OK, if you have 'special' request they will turn out not to be so flexible.
Why they are so cheap: no clue, to be honest. They are somehow related to the BGI in china (spin-off?) perhaps they are "sponsored" in some way?
Thanks for your feedback. I'm assuming you're speaking from personal experience, or from colleague's experience right? The website and the data they provide seems pretty good, but I'm wondering if someone here actually worked with them. (It sounds like you did).
And yeah, I'm planning to do a standard gene expression analysis, so it should hopefully be pretty straightforward.
I analyzed the data from Novogene recently. We sent the same samples to them and also sequenced them in house, then compared. Novogene looked VERY different. PCA was all messed up and didn't look good at all. We had immune cells if it makes any difference. Decided to stick with our sequencer, maybe it just doesn't work for our samples.
I don't know anyone at the company nor used their services, but guessing from their site:
they have lots of big, cutting-edge sequencers, which produce cheaper base per dolar than the sequencers available at most labs / cores.
with so many sequencers, maybe they also can get discounts at reagents / kits.
depending on where you are located, reagents and machines are a lot more expensive than in the US and / or China - this is the case in Brasil. Also, reagent orders take longer to deliver, sometimes creating a backlog, and diminishing reagents shelf-life - many Illumina reagents have somewhat short shelf-life.
If the list of publications is to be trusted, then yes, it seems they provide a good service. I've found many papers using their services, searching google scholar.
Based on the quotes they sent us last time on 20M/PE150, we estimated roughly $67/G, including library constructions. Quotes from BGI are about $65-80/G, depending on how many samples per lane. While promotional prices may make one company cheaper than another, IMHO, differences in price are negligible. BGI has been working very well for us, especially their turnaround has improved drastically in the last couple years.
I just got back 38 samples of ChIP~seq data; many reads in each but the mapping was completely wrong. Clearly there was something wrong with what they are doing.The rep seemed to be excellent but he did not tell us that they often inform customers that their samples have failed QC when they have been sent samples that are perfectly OK. For example, we send them volumes of 25 microlitres and they tell us that there is only 15 microlitres in the tube. If they cannot even measure volumes accurately how can they provide any reliable data? I second jiahaohaox above.
ChIP-seq is very difficult. I'd do this over several cycles in house if possible first to test antibodies and workflow. At our core we do multiple iterations with the customer before we get good results, and then we go to a bigger sequencer for higher throughput.
I would never recommend just sending samples out for ChIP-seq externally without having a lot of local experience with both wetlab and data analysis.
I understand that this thread is tempting to let of some "steam of frustration" (and yes ChIP can be super frustrating, I know well by experience) but I still encourage everyone to stay somewhat professional. Please give details and explain circumstances. Eventually this may help to decide if the company is the problem or if the error might be on your side.
If the company claims that little volume was send, then please provide details if you indeed sent the samples in a evaporation-free tube such as a Safelock and if you properly sealed the tube additionally with a lot of Parafilm.
Please indicate if you properly QC-ed the sample. Why were the samples "excellent"? Did you do a shallow sequencing or qPCR before to verify? Or did you base this statement on DNA concentration and behaviour on a gel/Tapestation/Bioanalyzer? How did you sent the samples? Dry-ice or ambient? All these factors can play a role beyond of what the company does to the sample. We always do a shallow sequencing with ChIP-seq samples before going for deep sequencing, even when using antibodies we use for years. Sometimes (and I do not know why) the IP simply does not work, be it the LOT of the antibody, any other reagent, my personal incompetence or the wrath of the ChIP god, I do not know, but it happens.
How do you know that many reads in each but the mapping was completely wrong? Is the sequencing quality bad, or the alignment, or can't you simply call peaks due to a lot of noise?
Thanks, ATpoint, In reply. I had checked everything by qPCR and by qBIt. I considered the samples excellent because the purifications were consistent with what we had done before and had been analysed by another genome centre that had produced very reproducible results over in three large and independent analyses. I sent the samples on dry ice. The antibody was an Abcam anti-GFP antibody and the ChIP was on yeast cells so given the small genome size it should have been easy. We knew what to expect because of the fact that the previous work had always produced consistent data. The new data gave very high peaks over sites that were different from what we had previously seen and which were different in different samples. The data was also largely inconsistent with the results of similar experiments from other groups. I conclude that ChIP~seq is tricky because of the small amounts of DNA and does require a dedicated service that one knows works well. I have asked for a refund. The company have been professional and have investigated my complaint in good faith but I remain disappointed with the waste of a lot of money and many months of lab work. There was one mistake in the previous post. I sent 25 microlitres and they measured 15 (not 5 microlitres).
Thanks everyone for the feedback , I m weighing the decision of sending my precious samples for shotgun metagenomics to novogene or not.
My issue is that I have had very very very very bad experience with bgi. my previous lab worked extensively with them and I cannot stretch enough how bad the communication was:
example 1) I had sent 200 16S PCR products to do 16S miseq sequencing and I had personally amplified and verified that what I sent was a 250bp region but apparently when it arrived in China, my pcr products were magically transformed to 1600bp size...
example 2) my colleague sent RNA from a fungal strain cultivated in vitro and the sequences he received were from shrimps, the story ended when my colleague showed them furiously the petri dish the RNA originated from over skype and yelled this aint shrimp...
(this is not a joke - this really happened!)
This is why the past few years I have favored companies in EU or US but now I have to consider novogene cause my collaborator is in Asia and he cannot send money for sequencing to Europe or US...
Therefore any experience with Illumina shotgun metagenomics of highly complex communities is very very welcome
Thanks for your feedback. I'm assuming you're speaking from personal experience, or from colleague's experience right? The website and the data they provide seems pretty good, but I'm wondering if someone here actually worked with them. (It sounds like you did).
And yeah, I'm planning to do a standard gene expression analysis, so it should hopefully be pretty straightforward.
data from colleagues / collaborators indeed, not data I personally had to work with