don't put the cart before the horse
I think some good suggestions have been made for narrowing down how XMRV operates in CFS and some other diseases, once we know that XMRV is in fact a causal element, or even a co-factor in CFS. But the Science article, Dr Bell, and pretty much everyone involved with WPI has stated clearly that their XMRV finding is tentative, preliminary, and needs follow-up.
I believe replication studies right now should address the basic weaknesses of the WPI study and determine whether or not XMRV is a factor in CFS. That is not yet proven. Speculation about XMRV makes a lot of sense for CFS, but that type of logic is not the same as empirical confirmation.
The Science article specifically raised several issues that need to be addressed. Additionally there are some obvious issue based on what one sees in similar studies.
Here are a few replication goals that come partly from conversations I have had with a retroviral expert. These are my words, not his, but I believe I have captured the essence of what he said.
1. Basic replication is essential. Several independent labs would use their own tests and collect their own CFS blood samples. There are over 20 types of PCR and many other types of tests that could be run. Just reproducing the WPI result this way would be a first step, often this type of scientific announcement falls apart at this first step, basic replication.
2. If WPI used a subset of the CFS population there is significant risk for the XMRV finding. A replication study should characterize the sample traits better. If XMRV is only in one subset then it can not explain all CFS as WPI is suggesting it might. Conflicting statements have been made by Dr Peterson (says they picked the sickest), and on the WPI website (says it was random from their 8,000 samples). Therefore, we do not really know if this is a finding of XMRV as yet another co-infection in a subset, or if it applies to all CFS. A replication study could stratify their samples and present data showing infection rate at different levels of CFS severity.
3. WPI made a statement early in the Science article that there was risk of contamination from MLV type species sometimes found in media. Although they believed they controlled adequately for this, not all agreed. So that issue must be addressed in a replication study. A replication could use more internal controls for contamination by other MLV type species that might share some XMRV DNA. They did sequence two samples, which addresses contamination risk, but that is not enough to be statistically valid.
4. WPI used an MLV type of antibody for their study, and if I understood that Science article properly, might not have an XMRV specific antibody. The 95% positive finding claimed by WPI might actually be MLV type antibodies, not specifically a finding of XMRV. Therefore replication studies should look for an XMRV antibody, or if they find only MLV type antibodies present in PWC, should throw out that conclusion of WPI.
5. WPI noted that XMRV might be a passenger virus. A replication study should look into that possibility. For instance, some immune bio-markers should be analyzed, and data correlated to XMRV infection rate and CFS severity levels. There are probably ways to analyze the data and give some probability that XMRV is a passenger virus. In other words, if XMRV is only present in CFS patients with collapsed immune systems and not found in most ordinary PWC it is probably a passenger virus. An unbiased CFS sample would be needed for that.
6. WPI has noted in the follow-up talks and interviews that XMRV may be a gatekeeper, lowering some immune functions and allowing co-infections to proliferate and cause CFS. This can be tested in a follow-on study, and really should be. To test for 'gatekeeper' status a study would have to be designed that compared XMRV infection rates in PWC with their own co-infection rates. For instance, do those with high XMRV viral load have more co-infections, particularly co-infections that might explain CFS symptoms? In my opinion, this is the most important replication study because it would present a possible causal model for CFS from an XMRV infection.
Several types of replication studies need to be run before clinical trials will make sense. Starting clinical trials based on unconfirmed research is a really, really bad idea, because then if you have success OR failure, you can not be confident of the reason why.
I think getting emotionally invested in XMRV as an explanation of CFS at this time is premature. The emotionally liable CFS population should not be encouraged to expect a cure from XMRV treatment before these types of replication studies have been completed, and even then only after successful clinical trials.
Some PWC here are pouncing on the XMRV-retroviral idea as a vehicle for CFS advocacy and politics, before ANY replications have been completed. I hope that does not come back to haunt us if XMRV fails to meet expectations.
But even if XMRV fails to explain CFS, the Science article has helped to raise awareness of our illness in some irreversable ways. More researchers are now aware that CFS is a problem and that we lack a proven causal model. That is the type of challenge researchers love to pursue. Also, we may have other retroviral infections that could be treated with ART, some good might come from this situation even if replication fails.