Name: Christopher Wong
Title: Chief Scientific Officer for Biomarker Development, Genome Institute of Singapore
Professional Background: 2009-present, CSO, Genome Institute of Singapore; 2008-present, head, Biopolis Shared Facilities, Singapore; 2006-present, officer in charge, high-content screening facility, Biopolis Shared Facilities; 2004-2009, research scientist, microarray and expression genomics, GIS; 2001-2004, research associate, microarray and expression genomics, GIS.
Education: 2001 — PhD, biomedical graduate studies, University of Pennsylvania, Philadelphia; 1990 — BS, molecular biology and music, Beloit College, Beloit, Wis.
Adding to a bevy of array-based approaches designed to monitor the different and evolving strains of influenza A, Roche last week announced that researchers at the Genome Institute of Singapore have developed a PCR-based approach that relies on a NimbleGen array platform to resequence the entire genome of any influenza A virus in around 24 hours.
The methodology, first developed by the institute in 2003 to identify and monitor severe acute respiratory syndrome (see BAN 3/31/2004), uses an in-house developed PCR-based technique to amplify the full genome of an Influenza A virus. The genome is then resequenced using a custom designed NimbleGen array to detect new mutations or reassortments in the virus strain.
The approach can be used with samples collected by nasal swab or nasal pharyngeal wash. According to Roche, the method could enable researchers to more quickly develop diagnostics for any possible new variant of the virus, and to determine rapidly if a strain has become drug resistant.
To learn more about the new flu detection assay, BioArray News spoke with GIS' Christopher Wong this week. Newly appointed as chief scientific officer for biomarker development at the institute, Wong has been at GIS since 2001. Besides his work in pathogen detection, Wong is studying breast cancer biomarker development and high-content screening. Below is an edited transcript of that interview.
I saw that you were recently named CSO for Biomarker Development. What kind of responsibilities does that entail?
Basically, I have been a scientist at GIS since 2001, and now we are beginning to develop some of our tools for commercialization. My role is to look at different applications, and identify the appropriate technology, so that we can have a product that can be used in diagnostics or prognostics.
I am a biologist by training, but when I came to GIS, my work had less to do with biology and more to do with technology. I have a wide experience on a variety of array platforms, and I really cut my teeth on those when we were developing array-based assays during the outbreak of SARS in 2003, to see what GIS could do for Singapore.
So, I have been involved in setting up various technologies, such as our new high-content screening unit, and my job has a lot to do with bringing in new technologies, and optimizing them so that we can do real work at GIS. I am obviously spending a lot less time in the lab these days. For this flu project, I have been designing the experiments and things like that, but I don't do the pipetting nowadays; someone else does the pipetting.
Could you describe the flu assay you developed using PCR and the Roche NimbleGen platform?
This is based on previous work we had done in pathogen diagnostics. We had developed a PCR protocol using proprietary random PCR methods that get rid of the biases inherent in random PCR. What we have done is to apply this same method to prepare patient samples and then resequence those samples using an array.
We are interested in detecting how the different strains of influenza evolve over time. There are various drugs that are being used to treat flu, for example, that can cause the strains to mutate. Our array is designed to have the whole genome sequenced, and to identify where the new breakpoints are so that we can detect new sequences.
When we were responding to SARS, we developed this approach using the first resequencing chip that came out of Roche NimbleGen. We had also applied the same method to look at the Dengue virus, because we do a lot of research on Dengue at GIS. When the swine flu came along, this method seemed like a natural one for us to use. The only alternative method would be to use a second-generation sequencing approach. Our sequencing group is managed by someone else, so I am not working with that group, but I know that it requires a lot of computational resources to put sequencing data together to form contigs. That approach takes more time and computation than array-based methods.
[ pagebreak ]
Could you describe the NimbleGen array that you designed?
It is a resequencing array, so, basically, for every position on genome we have four oligonucleotide probes comprising one perfect match and three mismatch probes. We use comparative genomic hybridization to see whatever signal is strongest and strongest signal is the sequence. That way we can just read the results of the assay right off and you can get the information very quickly.
The problem with flu is that it is quite a diverse genome. It requires a lot more probe sets than SARS or Dengue to accomplish the same screen. The NimbleGen platform gives us a lot of flexibility. The first version of the chip we designed for flu had 300,000 probes on it. Now, we have ordered a newer version of the chip that will have 2 million probes on it, because we want to cover all strains.
Why did you require that extra content?
Well, the chip was designed originally for the new Influenza A H1N1 flu, but we didn't have any of the new strain samples to test it on. So, we used seasonal strains, and for one of the strains we had we didn't have much coverage. The result wasn't optimal, so we decided that we really needed to have comprehensive coverage of all the various genomes on the chip. We know that there is a lot of redundancy in the current array we ordered, but we needed to get something usable fast. Based on results in the future, though, we will likely cut down on the number of probes.
Is it a singleplex or multiplex format?
For the 300K chip, we are performing assays in a singleplex format, but, if we wanted to, we could order chips in a threeplex format and run three samples on one array at the same time.
How does your approach differ from existing array-based methods for identifying mutations in flu strains?
Tessarae, for example, has developed a flu-resequencing chip using the Affymetrix platform. [Tessarae’s RPM-Flu array does not determine the entire genome sequence of flu, but can determine a "significant amount of the genome sequence to detect new strains and determine reassortment," according to the firm — Ed.] From our knowledge, no one has anything similar to what we have designed. Others have done resequencing on arrays, of course, but for other purposes.
Will your method be used by the authorities to monitor influenza in Singapore?
Singapore has formed a task force on H1N1 that involves clinicians from all local hospitals and the local CDC, as well as infectious diseases experts from various institutes and universities. I am part of that group and involved in developing diagnostics and technology platforms. Now, we are studying all five H1N1 samples we have in Singapore. So, there is the expectation that this technology will be applied in the field. We used similar technology during SARS to verify where different strains came from. So we have a good history of collaborating with the Ministry of Health.
Will you seek to have it used clinically?
We are not trying to use this chip for diagnostics. You don't need to resequence a whole genome to make a diagnosis, so it is more economical to use an RT-PCR test. But, under the infectious disease task force, we so have access to samples. So, it is more of a surveillance deal right now. There are patients that are reacting differently to the virus, and we can look at the different viruses and see if there are particular forms that we need to worry about. But, no, we don't intend to use this as a clinical diagnostic.
What has been GIS' relationship with Roche NimbleGen, and what other array platforms does it use in its projects?
Basically, we are more content driven, and each platform has its own niches and new product development, so we go with what is best at the moment at the point of study. When I started at GIS, I helped Lance Miller set up a microarray group and we developed our own spotting platform. We did those for a while, and now we have Affymetrix, Agilent, Illumina, and we use NimbleGen. We also use microRNA arrays from Exiqon. We use whatever seems to be the best, so at GIS we have all the major platforms available.
Are you satisfied with the quality of the available array platforms?
Actually, yes. There is a lot of pressure on array platform vendors to produce better products, because the cost of sequencing has come down. Both the array and sequencing platforms have their advantages and disadvantages. One thing we are doing is keeping track of the new sequencing applications, and making sure that we are using best methods available.
You also have ongoing projects in breast cancer diagnostics development and high-content screening. Could you give me an update on both?
The breast cancer work is part of our biomarker development program. We have a set of biomarkers that can be used to diagnose different kinds of cancer. Breast cancer has three grades. Patients with grade 1 tumors have a good prognosis, while those with grade 3 have poor prognoses. Those with grade 2, though, are harder to tell. We have developed biomarkers that can separate those grade 2 patients into grade 1-like or grade 3-like categories. We have found that those who have grade 1-like tumors have better survival rates, while those with grade 3-like tumors have a poor prognosis. We feel that these biomarkers could be useful in clinical use. So, my role is to try and develop the methods to a point where we can easily make this technology accessible to clinicians and to develop an assay that they can use in a regular clinical workup of their patients.
In terms of HCS, since we have done a lot of array experiments, we have generated lots of biomarkers. HCS allows us to do cell-based assays in a systematic fashion to examine these biomarker lists. You can knock out or overexpress these genes in cells, using robots, and then stain them to see if genes have function or not in what you are interested in. My role has been to develop that program. It is going well. It took a year to set it up, and it it's just about to get started on some projects.