#cloud sequencing

Author: Gabor Bethlendy

Below are the top reasons. If you think they have merit then read further.

1. Reagent “discount” economics don’t add up.

2. There are already too many next generation sequencers (NGSs) in the world already, and buying another under-utilized sequencer is a business model that favors the NGS manufacturers, not the facilities who use them.

3. More instruments does not equate with more access.

Let’s start with #1: Reagent “discount” economics don’t add up

First, the high price of, and poor access to, NGSs combine to drive up their service fees.  Second, the full price of the instruments is being amortized across the life of the instrument (typically three years)—when only 48% of its capacity is being used during this period (48% is the median utilization rate)! These amortization costs are also wrapped into genomic service fees.  

So what do the clever sequencing vendors do?  They offer you a second or third sequencer for a large reagent discount arguing this will bring down your per sample cost and hence the cost of your services when in reality all of those costs have already been captured in their calculation of the instrument price itself!

This is a false economy, or put more simply, there’s no such thing as a free lunch…

Which leads to my next point…

#2  There is already too much sequencing capacity, and more sequencers merely serves the NGS vendor’s interests

According to the latest figures, Illumina accounted for nearly 80% of the market share of next generation sequencers and about 4,800 instruments globally. By my calculations, this is >1 peta base-pairs of global capacity (to keep the argument simple, we’ll ignore the other NGS providers like Life Tech, Pac Bio and Oxford).

This is enough to sequence about 1 million genomes at 30x coverage on one of the Hi Seq instruments.   At the moment there are nowhere near 1 million genomes being sequenced each year and genome sequencing only represents a small (~10%) fraction of the NGS use case.  So why are institutions buying more of them?

One reason may be the above-mentioned 48% median instrument utilization rate. So while theoretical capacity may be high, demand out strips capacity due to poor instrument utilization.

Figure: Worldwide sequencing overcapacity is reaching ridiculous levels

We are often lead to believe that core facilities and sequencers are “busy” because we hear (or experience first hand) how long it takes to get on an instrument and get data back.  The erroneous conclusion we reach is to buy more sequencers to get better access, faster access or better data quality.

This is not the full picture. Yes, we often hear core facility managers argue that this is because NGS is a “very complex” workflow or that it is “not a commodity” and therefore difficult to scale up.  And while I agree in part, the fact is that academic and even many commercial labs are very bad at making full use of the sequencers they already have.  Either they don’t have enough business or more likely that their mission is to make discoveries not to make things more efficient. 

However, these are all truisms that distract us from the huge problem of poor access to expensive capital equipment in all of Life Sciences! The problems are efficiency and access.  And, the fact that these concepts aren’t as a sexy as that paper in Nature or Science is used to full advantage by the sequencing vendors; they are quite happy with the status quo as it keeps us naive scientists writing grants with taxpayer dollars merely to improve NGS vendor profits.  

Which leads me to the last main reason not to buy another sequencer…

#3: It’s all about sequencing access!

By having us buy more instruments for a larger reagent discount the vendor has done a Jedi mind trick on an entire industry all to give us “better access” and “some savings.” And as we discussed above, this is a false economy.

So, is the main problem the cost of sequencing or better access?  

I know many of you would say; “but I have a core or I send my samples to a collaborator or I search Google” and it’s easy for me to get on a sequencer.  And I think this is what most people said before Uber and ZipCar…that it is easy to hail a taxi as they are all over the city.  Yet, today Boston alone has about 40,000 Ubers (not to mention Lyft) circling the city on a daily basis!  How is this possible when there are enough cars and public transportation available already?  

Uber and ZipCar have lead consumers to an important understanding…for most, transportation is a service, and like many services these days, it is often less expensive and more convenient to rent it when we need it.

I am an American, and having my own car was really tied into my notion of American freedom and independence. But there is one thing I am even more attached to—my money.  That and anything that can lead to less headaches or more time in my life.  The convenience of ordering (and not owning) transportation has certainly helped in this regard.

Today, in the Life Sciences we are just like many car owners before Uber and ZipCar…we do not yet appreciate how difficult it is to get access to sequencers and other instrumentation…nor do envision how convenient it can be!

We have accepted the pain of how it is today.  

We have accepted that we send it to the local core or our favorite service provider.

So here’s my question: Assuming quality were the same, would you care which sequencer access you use? One selected through an Airbnb-like application or would you rather wait for your “favorite” one like you do today?