Skip to main content
Premium Trial:

Request an Annual Quote

UCSF s Joe DeRisi on Microarrays in the Hunt for the SARS Virus


At A Glance

  • Joe DeRisi
  • Assistant professor of biochemistry and biophysics, University of California, San Francisco.
  • BS — University of California, Santa Cruz, biochemistry and molecular biology
  • PhD — Stanford University, biochemistry.
  • Research interest — The DeRisi lab works on three areas: yeast molecular biology; the biology of Plasmodium falciparum, the organism that causes malaria; and viruses. Has NIH funding to study malaria and private funding for other work.

How did your collaboration with the CDC come about?

We contacted the CDC when this thing first hit the news [to say] that we had this virus chip and offer our services. They sent us samples — as well as many other people. We took the samples, which were just simply nucleic acid on the nanogram scale in many cases, and applied them to our virus microarray. About 24 hours later, we knew what was in the sample and we reported our results back to the CDC. The results indicated the presence of a novel coronavirus.

When did you make the first contact with the CDC?

Well, e-mails were sent, phone calls were made. The hard facts are that the samples arrived on Saturday, on March 22. FedEx delivered them at around 10:30 am and I was waiting by the door [with] Dave Wang, a post-doc in my lab, and a graduate student.

What did you do with the samples?

We received no live virus — we didn’t want to deal with that. We received total nucleic acid, DNA, and RNA all mixed together. We then amplified that mixture. We just labeled it [with] Cy5. We have a special probe for the array that is in Cy3. Then we applied samples to the microarray at about 9 p.m. or so. The next morning, at around 8, we scanned the result on our Axon 4000B, which, by the way, is the only piece of commercial equipment we have for this process; and then we read the results and sent them right off. We used GenePix Pro 4.0, which is the only piece of commercial software we use.

Was this a novel use of a microarray?

From a global emergency rapid response standpoint, I think it is a first for the array field. It’s a global emergency, they send you samples: 24 hours later, you send them results. We were not the only ones to identify coronavirus. [Others] identified it by EM [electron microscope] and also by degenerate PCR. So, we are an independent line of evidence. That’s what we designed it for. It was not just a fortuitous coincidence that we happened to have an array that could do this. We’ve been planning for this; in fact, we speak of this sort of use in that article in PNAS.

How long has the array been in development?

We’ve been working on it for two years. A postdoc in my lab, Dave Wang, who was first author on the paper, prior to working in my lab, worked on Kaposi’s sarcoma-associated herpes virus, along with [UCSF Howard Hughes Medical Institute investigator] Don Ganem. Kaposi’s sarcoma herpes virus is associated with disfiguring lesions [in] immunocompromised people, especially people with HIV. It was a mystery what caused this, and it was Don Ganem’s lab that discovered that it was a novel herpes virus. That sort of paradigm for virus discovery, and the paradigm that Chiron used to discover hepatitis C, is something we wanted to do — on a massive scale. Instead of one virus at a time, looking for one thing in an arduous way, we wanted to create a massively parallel system for viral discovery.

Will you describe your array for me?

It’s called the Virochip. There are 12,000 individual viral oligos on the array, and we will be expanding that to 20,000 by the end of the year. The array contains every completely sequenced virus from GenBank. We designed it based on a taxon-by-taxon approach. We have about a thousand taxons of viruses. There are hundreds of papilloma viruses that have been sequenced, and they all vary by just a tiny little bit. Obviously, we can’t put down every tiny variant of every papilloma. Human papilloma virus is one particular family and we designed a prototype set of oligos based on that model.

We use 70-mer oligos, synthesized by Illumina in San Diego. We use them exclusively these days.

How are you using this array other than this one event?

We are involved in the search for novel viral pathogens — causing hepatitis, presumed encephalitis — and many other diseases where there is a strong suspicion of a viral etiology. That’s the idea; you don’t want to take just one disease and hammer on it. You want to go for dozens of diseases. You want to maximize the possibilities of finding new viral pathogens. That’s where the interesting biology is.

When I first called you, your voicemail was full.

My voicemail is always full. The SARS thing has brought us some publicity, and that’s nice, but it’s not a big deal.

And this was all done on your own equipment?

We are sort of a do-it-yourself lab; that’s the way we have always done it. We built our own arrayer, a custom high-speed machine, and I wrote my own software to drive it. I have been doing that since 1996, so it’s no big deal. We make our own slides and print via mechanical deposition with our own printing tip design. Major Precision in Arizona does all of our machining. We have total control over every aspect of the entire process, and we don’t rely on anyone outside the organization to provide something.

Most people are limited by what arrays they can buy in a catalog, from a limited number of vendors. We wanted to do our own novel microarray, make our own oligos, of our own design. The only way we could do that was by containing all parts of the process.

Does managing this whole system take a lot of time?

It’s part of what we put our time into, and we are glad to do it. Because no one can do it like we do it. That’s the bottom line.

How could someone else replicate what you have done?

I teach a course at Cold Spring Harbor [Laboratory], and have for four years. I’ve now moved that [course] to California. It’s going to be part of the QB3 (Quantitative Biology 3) consortium between UC Berkeley, UC San Francisco, and UC Santa Cruz. As part of that consortium, we will be running courses, much like Cold Spring Harbor does. It will be taught this summer in August. My colleague, Vishy Iyer, is going to continue teaching the course at Cold Spring Harbor.

We’ve always had a strong teaching commitment to helping people build their own microarrays. For example, we put out a guide on how to build an arrayer many, many years ago. And literally dozens and dozens of labs have built their own arrayers and printed their own arrays based on our protocols and directions. We also run a website called, a clearinghouse for protocols. We do that out of our spare time, there is no money for that, there’re no commercial endorsement, no sorts of external funds whatsoever. We have some discretionary money, but we are not using any federal funds. We are just using the extra capacity that we already have.

Did you receive any funds for the CDC work?

Oh, no, no. There’s no money for that. Obviously, having a successful situation like that will help us apply for grants in the future; that’s really where the payoff is.

Can you compare microarrays to PCR and electron microscopes for virus hunting?

With PCR, you only get what you look for. There are literally thousands of viruses that one could look for, so you have to have an idea of what to look for before you do PCR. Our arrays are geared for the situation where you have no idea what you are looking for. Now, the CDC did their PCR based on results that they got [from] electron microscopy. They only knew to look for coronavirus when they saw what looked like a coronavirus particle in the EM. It’s fortunate that coronaviruses have a very distinct look; that’s a very strong tip-off. Now, on our samples that we are doing now for our clinical investigations, we may have only tiny, tiny amounts of samples from nucleic acid that don’t represent material that could be put into an electron microscope. And in fact, the virus, if it is there at all, may be present in such small quantities that it may escape detection by electron microscope. And, also, EM is only good when the particles are very, very distinct and can be distinguished from other types of viruses. So, I would say it is a complementary approach and it has several advantages in terms of time and efficiency.

Is there an issue with cross hybridization on your arrays?

The way we designed the chip, we try to maximize cross hybridization on the chip. That may seem counterintuitive, but remember, when we are looking for a novel virus, we want to maximize the probability that we detect it, so we selected regions of viruses for representation on the array that are the most conserved nucleic acid sequences among the families of viruses. By far and away, cross-hybridization false positives are not your worry at all. Your worry is that you get nothing at all. Nucleic acid hybridization is incredibly sensitive and very specific. And, if you have a novel virus, what is the chance you are going to hit precisely the right 50 or 60 nucleotides that you require to get a signal? The probability is very low, especially if it is a virus that has never been seen before. So, we really want to maximize the probability that you have any homology at all. If it’s a known virus, we are going to get a signal. But, if it’s an unknown virus, its really a gamble. But we are trying to stack the deck in our favor by doing very careful selection of the oligos that we put on our array.

There is still some discussion over whether this is indeed caused by a coronavirus.

Nothing that we did establishes causality. We only identified what was in samples sent to us. It is up to the CDC, in a much broader investigation developed from this, to try and establish causality. And that is pretty tough. It involves monitoring patients to see if they seroconvert during the course of infection — looking for antibodies produced for this particular antibody during the course of infection and verifying that essentially everybody fits the clinical patient profile. But the CDC [researchers] are experts in doing this. The interesting part now is sequencing the entire virus end to end. The CDC will do that.

Will you continue to be involved in this?

I doubt it. Our role in this was to identify the family of the virus and relate that to the CDC and let them run with the ball. It is not our intention to work on SARS in any way. It is not part of our current funding paradigm or the grant that we have written. It is just fortuitous that we were able to help.

What would you like to see improved in the technology?

We are working on this new technology to further and extend our viral microarray. It may not be completely obvious, but we have now developed the technology to recover sequences from individual spots that have hybridized. That is a major step forward.

How so?

Well, I’m not at liberty to say that until we have our preliminary disclosure to the OTM folks [Office of Technology Management] here. Abstractly, you want to recover the sequences from the individual spots on the array. No one has done that and it’s really obvious. It will take a couple of weeks to implement in the lab. The idea is that if you get a novel viral signature, the next step in deducing what the virus is, is getting sequence. How are you going to get sequence? Your array is already a platform for purification of those viral sequences; now you just need to go in there and pick ‘em out.

The Scan

Researchers Compare WGS, Exome Sequencing-Based Mendelian Disease Diagnosis

Investigators find a diagnostic edge for whole-genome sequencing, while highlighting the cost advantages and improving diagnostic rate of exome sequencing in EJHG.

Researchers Retrace Key Mutations in Reassorted H1N1 Swine Flu Virus With Avian-Like Features

Mutations in the acidic polymerase-coding gene boost the pathogenicity and transmissibility of Eurasian avian-like H1N1 swine influenza viruses, a PNAS paper finds.

Genome Sequences Reveal Evolutionary History of South America's Canids

An analysis in PNAS of South American canid species' genomes offers a look at their evolutionary history, as well as their relationships and adaptations.

Lung Cancer Response to Checkpoint Inhibitors Reflected in Circulating Tumor DNA

In non-small cell lung cancer patients, researchers find in JCO Precision Oncology that survival benefits after immune checkpoint blockade coincide with a dip in ctDNA levels.