Hi, everyone! For the idea, you can skip to the fourth paragraph.
Being bipolar, I usually work in fits and starts- nothing for a while, then all of a sudden a ton of ideas and inspiration come to me and I will have the energy to sit and work on them, and I can actually manage to get quite a lot done. My coworkers and former employers can tell you that I will go through times where I send spates of emails with new ideas, and these sometimes get passed over because, well, a brand new idea is a lot to deal with.
For my environmental studies, I had a couple of ideas in this current spate of productivity: to take some reports I had written a year ago or so and use them (the ideas in them, which are mine) to write methods/review articles describing the analyses I did and using the data I analyzed as an example of how to do it. I personally think the data, which dealt primarily with a tillage experiment, is worth publishing on its own but it would be impossible to make solid claims (too few data points), so I think that I’ll wind up presenting it in the context of being an example data set in a methods paper (on ecological network analysis, and another paper would focus on some more abstract ideas I have for looking at things such as functional redundancy in ecosystems using network data, and probably public datasets).
So that’s good. But I’m most proud of the idea I had today, which evolved over the course of the day as I thought about how to use existing data on the probiotic I studied in 2014, Lactobacillus johnsonii 456. I don’t know what stage clinical trials are at with this probiotic, but as the day progressed, I realized: this idea can be used for ANY CLINICAL TRIAL that is sufficiently advanced to have a large number of patients.
Here is the idea:
1. Clinical trial with lots of people: treatment, no treatment.
2. Take samples or use archived samples to determine if patients were infected with the virus.
3. You now have four groups: treatment, no treatment, both with and without viral exposure.
4. Statistical analysis to determine if the treatment can reduce the probability of infection and/or reduce the probability of experiencing severe effects of viral infection.
So instead of starting a trial now and hoping that the drugs studied will work, and waiting 12-18 months for a result, we can take a look at everything that’s being studied now which is close to ending. The wide spread of the COVID-19 virus means that a lot of people enrolled in most trials could have been exposed, and it’s not unreasonable to assume that there may be enough to make statistical tests valid, assuming that we can assay patients for exposure to the virus adequately. It may be possible to use epidemiological data and the residence of each patient to determine the probability of when patients were likely to be exposed, if this is required in determining effects. It should be possible to track symptoms, especially for patients requiring hospitalization. And if patients are monitored from now on with this analysis in mind, if they do become ill they can be watched to see what happens.
It’s not ideal, but it is a potentially effective way to leverage a large amount of existing work to rapidly find treatments which can help protect (or alternately, increase risks) from COVID-19. Current efforts to contain the virus are not sustainable- and in the case of the USA, there simply is not enough medical equipment to be able to treat cases adequately. While the government struggles to find a response and states also are doing their best, the research community can step up, and see if their clinical trials can be used to assist in the race to find treatments, or identify additional at-risk groups which may not yet be considered.
Update: I heard from Dr. Schiestl, my UCLA collaborator for the stuff I did in 2014, or someone at his address (I struggle a lot with wondering if these emails are actually from him- yay, bipolar) and he is supportive of me writing up a small paper describing my findings (recent ones, based on what is essentially sequence matching) with data I collected in 2014 from Lactobacillus johnsonii 456, and public database human proinflammatory cytokine genes (involved in the cytokine storms which cause lung damage and organ failure in severe cases of COVID-19). I figure this is important data, and it builds on my general hypotheses about how bacteria interact with eukaryotes, so I’m just going to submit that hypothesis paper (already written) and this new set of analyses involving proinflammatory cytokines. I’ll submit the hypothesis to biorxiv, and this followup relating to novel coronavirus in medrxiv. I’ll post when this is all ready to go. There will likely be problems with my analyses, because this is not really my area of expertise, but I’m very happy to inspire people to look at my data and their own data, and consider the novel ideas I’ve had about how bacteria interact with us, particularly in the context of fighting off this pandemic.
Update #2: papers submitted. It will take a little time for these papers to clear the review process (not peer review, which I intend to seek once I figure out who might want to publish really highly speculative papers, but just suitability for the archives). [Note: both papers rejected for not being full research articles, keep reading.]
I had a hard time figuring out what category to use for the submission to MedRxiv as “microbiome studies” is not a category on MedRxiv, and that’s basically it- plus, the clinical trial idea is possibly more important than the things I wrote about probiotics or dysbiosis, and that idea is really why I bothered to submit these papers now at all. I’m hoping all of both submissions will be useful, but I believe even if all my thoughts about bacteria are not, at least this one idea about trials should be worth mentioning and circulating. Once the papers are in the archives (fingers crossed), I’ll post links and start mentioning the clinical trial idea in letters to medical journals, since my attempts to reach clinical trial companies, health agencies, and newspapers have failed. It’s also possible that people have already thought of the clinical trial idea, but if they have, no one is talking about it.
Update #3: Well, this is a wrinkle. Submission to BioRxiv was rejected for being a hypothesis paper, not a full length research article. Too short. It’s so hard to publish a hypothesis with only preliminary data, and I haven’t been able to test the hypothesis myself. Every time I try to get a US lab to open a Grants.gov portal so I can write an NIH grant, for example, I get no assistance, and I can’t do work with UCLA mice and a UCLA bacterium and a UCLA collaborator unless it’s pretty much going to be at UCLA. This has gone on so long I’ve basically given up. It might not even matter, if the other submission is also rejected for being short.
Update #4: Well, as I half expected, my MedRxiv submission was also rejected for being a proposal, not a full research article. I think my next steps are, if my agitated reply to the staff at MedRxiv is ignored, to try to write letters to the editor about the trials idea in papers like Nature and Science and The Lancet, and hope there aren’t expensive page charges for that or a lengthy consideration process and messing with paper formatting. My hypothesis paper might go to mBio, and I’ll just fold the work I did relating to proinflammatory genes and so on into the supplementary information for that.
I’m more concerned about getting this idea about using clinical trial data out quickly, though. People are dying. The stuff I wrote about dysbiosis and COVID-19 symptoms, which draws upon both papers and is sketched out in the second, might be helpful longer term but I think the trial data idea is most important. The clinical trial idea I might be able to publish in The Conversation, and hope enough clinicians will see it.
I’m very tired today- I think that the stress of this pandemic plus staying up late for 3 nights in a row repeating analyses and revising and writing has caused a flare up in my bipolar disorder, and I’m just going to have to take it easy for a couple days while I try to think of what journals most clinicians read, and figure out how much submissions might cost. It’s also occurred to me to try to write letters to newspaper editors, but again, not right this second, I’m pretty tired and I don’t know if those would be published, either, or circulated to the right people. I wish I knew someone prominent who would help me with this, I honestly think I’ve come up with a very good idea but I’m not getting any help in circulating it to the people who need to see it.
Even Dr. S. isn’t saying “oh, hey, let me use this idea on my trials that are about to end” and that is pretty strange to me (and it makes me wonder, again, if it’s not actually him I am communicating with, hooray). All you need to do is swab all your patients for COVID-19 and then split them into infected vs non-infected for each treatment. It would be perhaps a bit expensive but any molecular biology lab can do the RT-PCR, provided that RNA is extracted safely first. The RNA alone isn’t dangerous. If you wanted to inactivate the virus to make it safer to handle, just soak it in proteases first to destroy the spike proteins, kill the proteases with heat treatment (which will not injure the RNA), and then you can just do the RT-PCR, provided the lipids and other sample contents don’t interfere with the reverse transcription or PCR and you work fast to prevent RNAses from destroying your samples. I wouldn’t be afraid to do this myself in a regular lab provided I got the sample after protease treatment, and provided I had evidence that such treatment destroyed the virus’s infectivity. And if you wanted to make this process safer, just have the technician swabbing people put protease in the sample tubes and start the incubation for the enzyme as soon as possible after. That way only one sample handler is exposed to intact viral particles.
It occurs to me that if you wanted to look at dysbiosis, it’s not perfect but you could assay the swab nucleic acids (archived, probably) for 16S and ITS sequences and determine the abundance (with qPCR) or composition (with barcoding) of bacterial and fungal communities as well. This is, again, more money but it is possible.