HNHacker News
TopNewBestAskShowJobs

rgejman

672 karma · joined December 26, 2009

http://ron.gejman.com [ my public key: https://keybase.io/rgejman; my proof: https://keybase.io/rgejman/sigs/jXEYtNnNbzvxR-zVpC8bW7XBAdDKvlfIRh2qhvWA3w8 ]
submissionscomments
rgejman··on San Diego Researcher Crowdfunding Patent-Free Cancer Drug
The "bait" is the claim that a patent-less drug will be cheaper when it finally comes into use in the clinic than a patented drug (from your indysci campaign: "Releasing without a patent means the drugs will be cheaper and it will be easier to build on the work to make improved drugs or drug combinations. Releasing without a patent means expanded access to drugs in countries that can't afford extensive licensing and export agreements.")

But you and I both know that is not how drug development works. When you bring in a pharmaceutical company to fund the clinical trials (which are hugely risky) they will demand an upside. The upside will come if they (a) make a derivative molecule that they can patent (b) patent a use of the drug for an indication (c) package the molecule in a way that is patentable -- e.g. with an adjuvant.

The crowdsourcing page on Indysci reads as if the drug will eventually be cheap and readily accessible just because you have not applied for patent protection.

Another point: the indysci page misleads people about what drug patents means by comparing it to open-source software. As an academic researcher, you can work on ANY molecule, no matter its patent status, in an academic research setting. That means you can develop new uses for the molecule, make derivatives, etc. What you really mean is that patented drugs cannot be SOLD by anyone other than the manufacturer.

rgejman··on San Diego Researcher Crowdfunding Patent-Free Cancer Drug
The NIH mandates that all research it funds be placed in the public domain within 1 year of publication. http://publicaccess.nih.gov/

This does not go far enough, but it is a good start. It would be better if non-published research were also placed in the public domain, but this is logistically tricky for many reasons.

rgejman··on San Diego Researcher Crowdfunding Patent-Free Cancer Drug
There are many problems with the institutions in this country that give grants. However, crowdsourcing is not the answer. Nor is baiting people by offering them the promise of a "patent free" cancer drug. Full-term drug development, including clinical trials reaches into the billions of dollars. You're never going to crowdsource that. So even if your drug is wildly successful and nontoxic (of which the chances are tiny given that you have not yet done animal studies), you will to partner with a pharma comnpany to bankroll the rest of its development.
rgejman··on San Diego Researcher Crowdfunding Patent-Free Cancer Drug
We already have crowd funding for science. We do it as a whole society when we fund the various grant-giving agencies in this country. I cannot emphasize how bad of an idea it is to take funding decisions away from professional scientists and place them in the hands of untrained individuals, no matter how enthusiastic they may be nor how much money they are willing to throw away.
rgejman··on MTA Responds Citing Lack of “Infinite Change” for MetroCards
I really don't think this is a big deal. I've had the same MTA card in my wallet for over a year (and a previous one like it for a year before that). I just top it off with $50-60 a month and there's no issue.

I appreciate that this might be bothersome to some tourists, but my experience is that most tourists who buy MTA cards don't calculate exactly how many rides they're going to take in a several day visit to NYC.

rgejman··on The curious case of the cyclist’s unshaven legs
Unfortunately, they are usually also the best equipped to find flaws in the new work. True, this is not ideal, but it's also not ideal to send a paper to less expert reviewers.

All of which goes to say that pre-print publication is the way to gooooo.

rgejman··on The curious case of the cyclist’s unshaven legs
There are two problems, only one of which is addressed by duplication.

1. The original study may be plain wrong. Author 1 claims to describe phenomena X, author 2 can't find any evidence for X using the exact same methodology.

2. The original study is wrong because the methodology is bad, interpretation is incorrect, reagents are not specific, techniques are not clean, etc.

There's a strong argument to be made that it is more important to do orthogonal work which indirectly verifies the original results, as opposed to simply trying to reproduce the original work. This solves both #1 and #2.

The paper cited in the article is a great example. Gluing hair onto a plastic leg and putting it in a miniature wind tunnel is probably an exercise in bad methodology—but they might have still gotten the right result for their experiment. And that's useful because now we know that there may be something different about either plastic legs vs. real legs or miniature wind tunnels vs. big wind tunnels. This is a trite example, but it should be clear how this might be more important in other studies.

rgejman··on Medicare: Not Such a Budget-Buster Anymore
Yes, I did. I don't think you're interpreting the results correctly. The conclusion of the paper, in summary of all the evidence, is that "Even against the background of advances in the management of obesity-related chronic diseases in the past few decades, our findings suggest that adiposity, including overweight, is associated with an increased risk of death."

Please look at the figures, in particular figures 1 and 2 D/C. Among non-smokers, as the BMI increases from 25 there is an increased risk of death. You can see the trend clear as day.

rgejman··on Medicare: Not Such a Budget-Buster Anymore
This is just not true. Being overweight or obese significantly increases all cause mortality. Here's a study where they followed 527,265 people for 10 years:

http://www.nejm.org/doi/full/10.1056/NEJMoa055643#t=articleR...

"During a maximum follow-up of 10 years through 2005, 61,317 participants (42,173 men and 19,144 women) died. Initial analyses showed an increased risk of death for the highest and lowest categories of BMI among both men and women, in all racial or ethnic groups, and at all ages. When the analysis was restricted to healthy people who had never smoked, the risk of death was associated with both overweight and obesity among men and women. In analyses of BMI during midlife (age of 50 years) among those who had never smoked, the associations became stronger, with the risk of death increasing by 20 to 40 percent among overweight persons and by two to at least three times among obese persons; the risk of death among underweight persons was attenuated."

rgejman··on For ‘Big Data’ Scientists, Hurdle to Insights Is ‘Janitor Work’
Agreed. We need to talk about these issues as a community and come up with effective strategies to prevent data contamination. We've done this for web development and the result was a panoply of tools to validate form input. Now we need to do this for our back-end tools.

For instance, statistical programming languages like R, which generally operate on nice tabular data, should come with built in methods to enforce data validity. I should be able to tell R that I expect certain columns to only contain values from 0-1 and no NAs. This is an easy example because languages like R are somewhat dictatorial about how they want you input and process data, but one can imagine the same sort of methods built into the base libraries of general purpose languages.

On the other hand, maybe we need to be more rigorous on the whole about data validation. We accept that automated/continuous testing is an effective mechanism for preventing bugs. We need the same automated systems to check data files and flag them when there are issues.

rgejman··on For ‘Big Data’ Scientists, Hurdle to Insights Is ‘Janitor Work’
Interesting. I'd love to hear the details.
rgejman··on For ‘Big Data’ Scientists, Hurdle to Insights Is ‘Janitor Work’
Data cleaning/munging is a surprisingly pervasive problem in my field ('big data' cancer research). I spend a good fraction of my time hunting down bugs caused by dirty sources upstream of my scripts. One way to fix this is to become more rigorous about using tools and libraries that check and enforce data correctness and consistency... but it's hard to do this on an individual level. Even if I'm careful about ensuring the correctness of my data, other researchers who send me data may not be as scrupulous. As with any challenge that requires teamwork and constant vigilance (e.g. unit testing), perhaps the data science community needs to invest in tools and computational frameworks that constantly monitor data for correctness and consistency using DDLs and the like.
rgejman··on Bikanta’s (YC S14) Tiny Diamonds Find Cancer Before It Spreads
Agreed, for now. Caveats TBD. In particular issues of access (e.g. astrocytomas, non-metastatic bladder) and identity (what does it mean to be a CTC? A known driver mutation? We know that some tissues are littered with such mutations...)
rgejman··on Bikanta’s (YC S14) Tiny Diamonds Find Cancer Before It Spreads
If they are in fact small enough to escape via biliary excretion, that would be great. But just skimming the literature it looks like they can accumulate and persist for weeks in the liver—a situation which would suggest hepatotoxicity is another problem (Zhu et al 2012 in "Theranostics" -- never heard of that journal but paper looks 'ok').

Coatings won't help this problem unless you can somehow change the shape of the crystal to fit through the slit OR get it to cross into the renal luminal space via transcytosis or excreted more efficiently in the biliary tree. All difficult propositions.

I really hope novel techniques like this work but this particular on seems very problematic. If I were you I would complete a serious set of toxicity studies very early on to make sure this isn't a fish that's dead in the water.

Edits: clarifying the points

rgejman··on Bikanta’s (YC S14) Tiny Diamonds Find Cancer Before It Spreads
I am very pessimistic about the likelihood of this being a non-toxic visualization strategy in mice or men. These "microdiamonds" are unlikely to clear the kidneys, leading to kidney toxicity. This is the problem with all ≥ nanoscale molecules, with the exception of very long ones such as nanotubes which can pass longitudinally through the kidney slit diaphragms.

Many a great drug (and visualization agent) has failed because of kidney toxicity.

It's surprising that there are no MDs on the team.

rgejman··on The increase of genetic complexity follows Moore’s law (2013) [pdf]
The idea that DNA length is costly to the organism is a venerable, but increasingly controversial theory. The existence of disproportionally huge genomes belies the hypothesis. While it's possible that some organisms derive benefit from having larger genomes, it's also reasonable to infer that most eukaryotic cells are not limited by energy. So much of eukaryotic cellular activity seems energetically "wasteful," yet cells just don't seem to care.
rgejman··on Combatting illegal abuses of ngrok
OP:

I'm sure the service you offer is great. Moreover, I'm sure that not requiring any identifying information greatly improves the experience you give to first-time users and increases conversion. However, you are inevitably doing a disservice to your users and the entire internet by offering hosting without any sanity checks (like an account, valid credit card or deposit). I hope you'll reconsider whether increasing legitimate usage of your service is worth inadvertently assisting criminal activity.

rgejman··on The Deadliest Animal in the World
Cows ought to be on that list as well. They kill 10-20 people per year in the US alone. http://tierneylab.blogs.nytimes.com/2009/07/31/dangerous-cow...
rgejman··on Immunity Project – HIV Vaccine Development Program [pdf]
I think that they will incubate CD8 and CD4 T cells from the same mouse together. The CD8 (cytotoxic) T cells, if they recognize the antigen, should kill the infected CD4 cells.

A caveat is activation/co-stimulation. The CD8 T cells should already be activated due to the vaccination and no longer need co-stimulation from DCs... but this depends on the vaccine having delivered enough peptide to stimulate a true immune response in a human/humanized immune system. That's the (first) big unknown here.

rgejman··on Immunity Project – HIV Vaccine Development Program [pdf]
That's an interesting approach.

Re not knowing if the human analogs are immunogenic yet: An alternative way to to answer your question would be to identify a T cell receptor (TCR) clone that recognizes your peptides of interest. There are several ways of identifying TCR clones that match your peptide of interest. Probably the best way is via phage display. The advantages here are manifold:

1. By identifying a human TCR you immediately prove that your peptide is recognized by the human immune system (immunogenic)--you still need to do a bit more work to show that the immune system will produce this antibody in response to your peptide--but that's a pretty good start.

2. You could perform phage display separately on TCRs from elite controllers and more susceptible patients. The differences in these TCRs would likely make for a very interesting academic publication.

3. You could make a TCR-like antibody based on this TCR. If your vaccination strategy failed, hey at least now you have an antibody you can take into trials!

rgejman··on Immunity Project – HIV Vaccine Development Program [pdf]
#1 suffers from the critique I mentioned above. The peptide(s) you are using are designed for H2d. Do you have evidence that their human analogs* are immunogenic?

#2 I assume now you are using the human version of the peptide? I'm not sure why you're doing this in vitro when you can now do it in vivo. Why not (a) infect the mouse with HIV, (b) look at viral titers w/ and w/o pre-vaccination, (c) look for specific CD8 T cells that have developed against peptide-HLA complexes.

I'm not sure I understand your in vitro experiment. Are you planning on mixing CD4 and CD8 T cells together in vitro from vaccinated/non-vaccinated mice and looking for increased killing of the CD4 cells (as estimated by decreased p24)?

*I assume it's mostly anchor changes here

rgejman··on Immunity Project – HIV Vaccine Development Program [pdf]
Another thing I want to point out is that your home page touts safety as a design point. The text under "SAFE" says "Our vaccine does not contain killed viruses, live viruses, or genetically-modified viruses." I STRONGLY object to this language as it implies that inactivated (killed) and live* viruses are unsafe in general. This is simply not true. I believe most clinically used vaccines are either inactivated or attenuated.

I understand that you're trying to make sure people understand that your vaccine is safe. However, implying that other vaccines are unsafe because they are inactivated or attenuated does a disservice to the entire vaccine industry.

*I think you mean 'attenuated' here, not live

rgejman··on Immunity Project – HIV Vaccine Development Program [pdf]
There are a lot of moving parts, so it's really hard for me to estimate the cost here. It's done somewhat routinely in some institutions (although it is very, very expensive).

You first need to make humanize mice (i.e. mice with human immune systems). Luckily we can make those by taking NOD/SCID mice (an [expensive] immunodeficient mouse line) and then colonize them with human hematopoietic stem cells from a human donor to reconstitue the immune system. Then you have to do a lot of experiments to make sure your humanized mice can mount immune responses. Then, you can test your peptide and see if your humanized mice will make CD8 T cells to your peptide... Finally, you can infect your mice with HIV (now that you have a humanized immune system) and see how your vaccinated vs. unvaccinated mice fare.

I have no idea what the downsides of this approach are because I don't work on HIV or with humanized mice--but mice ain't people. So after all that work you'd have shed limited light on what would happen in real people.

I would argue that a better approach is to see whether HIV patients have actually made CD8 T cells against these antigens. This can be done more inexpensively and would give you pretty good evidence that your peptide is immunogenic in real people.

rgejman··on Immunity Project – HIV Vaccine Development Program [pdf]
Some thoughts on this white paper, with the caveat that I'm not sure how much of what is discussed they have done and how much of it is what they are planning to do. The white paper is light on details and I re-iterate my appeal for them to release their data (at least in summary form). My apologies for some of the technical language--I want to write this first and then I can clarify the hard stuff (or you can google it)

This vaccine relies on inducing immunity to an HIV peptide* (i.e. a ~9 amino acid cleavage product of an HIV protein) that is predicted to bind to HLA molecules. At first blush, this sounds like a great idea--let's immunize people to the very thing that the immune system is supposed to recognize as bad! Unfortunately, this has been tried lots of times and usually doesn't work very well, if at all. It's been tried in cancer, it's been tried in HIV, it's been tried in practically any system you can imagine. I'm not saying that peptide vaccines don't ever work (hell, my lab has worked on one)--I'm just saying that they have had lots of promise and limited successes. In fact, animals and patients have been immunized to many cleavage products of many cancer associated antigens to little, if any, effect—or effects in animals but not in humans. We can hem and haw all day about what the specific studies show, but suffice it to say that peptide vaccines are an old idea. Perhaps Immunity will get lucky here, but I'd be surprised if there isn't a lab somewhere that has tried immunizing animals to every single breakdown product of HIV proteins already.

Immunity's white paper says that their putatively immunogenic human peptide cannot be tested in animal models because it will only bind to human HLA. So they are planning to use a mouse-analog of the human peptide to do their animal testing (or have already done this?). That is, they are planning to use (or maybe have already used?) a completely different peptide (probably 1-3 amino acids different from the human version) in animal studies than what they would use in patients. This may be a significant difference... or it might not be. There's no way to tell a priori. A more classical and IMO better approach would be to identify T cell receptor clones from patients with HIV (or without HIV) that bind to these peptides (using tetramer staining, phage display or some other high-throughput technology), demonstrate that humans are capable of creating T cells that recognize their putatively immunogenic peptide followed by T cell killing assays on cells presenting their peptide of interest. There are other approaches, but that is a commonly used one in the field.

* in some places the white paper says "peptides" and in other it says "peptide" so it's not clear if we are talking about 1 peptide or multiple. I think they are mostly talking about multiple peptides, because they want to immunize people no matter their HLA type. Different HLA types bind to different peptides. So if you have HLA A02 you will bind different peptides than if you have HLA A24... and so on for all the different HLA types.

rgejman··on Startup developing HIV/AIDS vaccine is 2nd nonprofit accepted into Y Combinator
I'm confused as to why you would encourage anyone to donate money to your organization without providing any data. I'm also confused about why you care about publication in traditional scientific mediums. You are starting a company and it would be in your interest to release as much data as possible about your successes in an attempt to raise money.

Not to mention, there are some journals that will publish pre-printed data. Here's a partial list (http://en.wikipedia.org/wiki/List_of_academic_journals_by_pr...) although take that list with a grain of salt because the pre-print policies may only really apply to math/physics publications.

rgejman··on Darpa to Genetically Engineer Humans by Adding a 47th Chromosome
This article is fear mongering with incomplete information.

There are huge problems with the current techniques we use to introduce DNA into cell lines of all kinds (human or other mammals). For instance, the virus techniques he describes for inserting DNA cannot control where the DNA lands in the genome. Thus, you may end up with a cell that seems to have the properties you are after, but it acquired those traits through disrupting or activating a normal gene, thus confounding your results. Alternatively, you may be comparing cell lines with viral insertions and would ideally like the expression of the inserted genes to be similar (so that you can fairly compare their effects).

These and other problems would be solved by having cell lines with known insertion locations. Whether this is best achieved by adding a new chromosome to these lines or engineering a precise "landing spot" in an existing chromosome is up for debate.

rgejman··on [dead]
Judging by her article history, she seems like more a tech columnist and not a science news writer.
rgejman··on Heat map of homophobic and racist tweets
A fundamental problem with this strategy is that is may bias the results towards prolific haters—i.e. people who are frequently hateful on Twitter. These people may live in places where social norms permit online hateful speech without repercussions. This strategy will miss people who are haters, but must be more subdued in their expressions of hatred online.

It would be interesting to identify people who have posted just a few hateful messages—perhaps few enough that they can get away with it in their local social context. This may more sensitive to occult haters.

Something like this: 1. Identify individual twitter users as being hateful in a particular category. For instance, user A uses the word "chink" 3x and "fag" 5x in 100 tweets, so he gets added to the "chink" and "fag" categories. Play with these threshold to see what makes sense. 2. Divide the # of hateful users in each category by the total # of users in that location. Allocation of users to location can be done proportional to the # of tweets they make from each location.

Cool project :).

rgejman··on Our Feel-Good War on Breast Cancer
I wasn't familiar with that. Is that stress-related? Mastectomy without chemo should not induce menopause, as far as I understand the physiological processes at work. If the patients are also being treated with tamoxifen or other estrogen receptor antagonists then the menopause is, I believe, reversible.
rgejman··on Our Feel-Good War on Breast Cancer
Just as an FYI: Women can also have some loss of sexual function, either physically (due to ligation of the 4th intercostal nerve that supplies the nipple), psychologically or both.
← PreviousPage 4 of 5Next →