Criteria for funding and promotion lead to bad science

[Darnok]

I've said this for years... and they had to make study to confirm such simple thing...

http://phys.org/news/2016-11-criteria-funding-bad-science.html

[Mikki]

@Darnok

Well, your comment and the lack of any reply (except mine here) confirms the major problem of current science funding, assuming most scientists need to work do research for a living, and the institution(s) which funds or pays said research, likes to see "specific" results and/or any possibilty of an economic exploit of said research.

Any otherwhise funded research by privat or company or country is not allways distinguishibly different in that practice.

The few remaining independent scientists or researchers of any branch remain mostly in secrecy and when one of them occurs occasionly with unconvenient results or other claims a general "public" sh*tstorm from "serious" well funded and promoted colleagues arises from the horizon.

Why? Money. Future. Fear. "Cheating Death". It is simply the point where reason is corrupted by fear.

History repeats itself, and the current human accelerated rate of spread around this planet doesn`t look really like there is "reason" in the winning, in fact, average humans become dumber any generation, the sheer amount of information available makes this place a huge mess running out of any natural regulation.

Good luck humankind, because luck is the only reliable scale for 99% of us.

Excuse my bad english. Thank you

Edited November 17, 2016 by Mikki

[KSK]

15 hours ago, Mikki said:

@Darnok

Well, your comment and the lack of any reply (except mine here) confirms the major problem of current science funding, assuming most scientists need to work do research for a living, and the institution(s) which funds or pays said research, likes to see "specific" results and/or any possibilty of an economic exploit of said research.

Any otherwhise funded research by privat or company or country is not allways distinguishibly different in that practice.

The few remaining independent scientists or researchers of any branch remain mostly in secrecy and when one of them occurs occasionly with unconvenient results or other claims a general "public" sh*tstorm from "serious" well funded and promoted colleagues arises from the horizon.

Why? Money. Future. Fear. "Cheating Death". It is simply the point where reason is corrupted by fear.

History repeats itself, and the current human accelerated rate of spread around this planet doesn`t look really like there is "reason" in the winning, in fact, average humans become dumber any generation, the sheer amount of information available makes this place a huge mess running out of any natural regulation.

Good luck humankind, because luck is the only reliable scale for 99% of us.

Excuse my bad english. Thank you

Except the article had nothing much to say about any of that. It didn't even mention any particular scientific field. It's main argument was that the current system favours small, exploratory studies which might generate new eye-catching results, rather than medium sized or large studies to confirm those results. No conspiracy theories in sight.

It seemed like a pretty fair and reasoned article to me. I would add a couple of things:

• Those medium and larger studies are also a lot more expensive so the chances of getting a research grant to pay for them is (sadly) correspondingly lower.
• The idea that scientists like to discover new things rather than repeating old work, doesn't come as much of a surprise to me - I would imagine that most scientists work in science precisely because they want to discover new things. On the other hand you also get scientists that care deeply about checking results and making sure they are reproducible - I know a number of them personally.
• I wonder if the model used in the article takes different lines of evidence into account? For example, scientist A comes up with a result based on a small study. Scientist B carries out a different small study which provides different evidence to support scientist A's result.

[Green Baron]

 

I'm not a follower of psychology, i must admit.

Well, there may be a spark of truth in it that the urge to bring out something thrilling leads to silly announcements (faster than light particles, alien stars with megastructures, reactionless drives, ....). I don't see any better way than the current principle of peer-reviewed publishing in scientific journals, in combination with hopefully fruitful discussions as a part of the process.

I personally too find the coming out on social media, magazines and news papers dangerous to the public perception of science, far too much nonsense is titled ("scientists have found out ..."), and on a closer look you only find hot air.

These are my criteria to judge the seriousness of a publication:

- the authors are actual researchers in the field

- it's published in a peer-reviewed journal without being announced on a pre-print server, social media, magazine ...

- it somehow leads to a discussion in the community, being cited by others, answered or asked directly

- if it's based on an actual experiment, that experiment can be reproduced without bending any laws

There is a grey zone, of course, but much stuff can be sorted out that way.

New discoveries are generally not done by a single person but a large team, in many cases from different scientific fields. It can take a few years from first finding until a serious publication is written. And during that time many things have already been discussed and adressed so that a thing like "we broke the laws of physics" is sorted out before it sees the light of day. On the other hand, a discovery that claims to break any laws, published on a social media network thing is likely to be bogus.

That said without stressing psychology :-)

 

Edit: we should distinguish between research for a given goal, mostly technological / engineering / medical, in any case connected with an economic thought in mind on the one hand and pure, basic fundamental research, in public institutions or universities, in some cases even done by people who choose their own goal, on the other hand.

 

Edited November 18, 2016 by Green Baron

[tater]

From the beginning of the article:

Quote

Some scientists are becoming concerned that published results are inaccurate—a recent attempt by 270 scientists to reproduce the findings reported in 100 psychology studies the Reproducibility Project: Psychology found that only about 40 per cent could be reproduced.

This latest study shows that we shouldn't be surprised by this, because researchers are incentivised to work in a certain way if they want to further their careers, such as running a large number of small studies, rather than a smaller number of larger, more definitive ones. But while this might be good for their careers, it won't necessarily be good for science.

I'd be more interested in the stats from, well, a real science. It's not all nonsense, but rather a lot of it is. 

Perhaps the title of the article should be, Criteria for funding and promotion in psychology lead to bad science (my emphasis, obviously).

If their findings can be generalized, then provide data for particle physics papers, or biochemistry, etc, and see how they compare.

I'm willing to entertain a pretty broad definition of science for disciplines using the scientific method, but the idea that they are all somehow equal is clearly rubbish. Some of the very, very soft disciplines out there will very likely become "harder" in the future, but right now, they're pretty sketchy. Any discipline with "science"in the title is usually pretty suspect (it reminds me of "democratic republic" within a country name, lol).

Edited November 18, 2016 by tater

[Green Baron]

16 hours ago, tater said:

*snip*

I'm willing to entertain a pretty broad definition of science for disciplines using the scientific method, but the idea that they are all somehow equal is clearly rubbish. Some of the very, very soft disciplines out there will very likely become "harder" in the future, but right now, they're pretty sketchy. Any discipline with "science"in the title is usually pretty suspect (it reminds me of "democratic republic" within a country name, lol).

Yes, the paper seems to address mainly psychological research, which is a discipline of medicine, right ? In general medical doctors must write a dissertation, so just the sheer number of medical dissertation makes it difficult for the authors to add something new to the discipline, usually the meaning of a dissertation. So medical dissertations are often statistical works which generally have the problem to choose an appropriate sample. Sometimes the sample is limited to a few dozen patients with a certain symptom or being treated with a medication that is in clinic testing. Here funding surely is a problem (who pays the study ;-)), but i think most are aware.

Imo this is a good point to check the seriousness of a publication in natural sciences. The scientific method includes the clear obligation to describe exactly the composition of a sample (where, why and under which circumstances exactly were the samples taken) and justify, if necessary, why e.g. runaway-data was dropped, etc. Usually this is done in serious papers, but to judge the correctness one needs a deeper knowledge of the case, often inter-disciplinary.

 

[YNM]

At least, for me, this news is pretty old. The phenomenon can be there, even in natural science studies.

Edited November 19, 2016 by YNM

[tater]

4 hours ago, Green Baron said:

Yes, the paper seems to address mainly psychological research, which is a discipline of medicine, right ? In general medical doctors must write a dissertation, so just the sheer number of medical dissertation makes it difficult for the authors to add something new to the discipline, usually the meaning of a dissertation. So medical dissertations are often statistical works which generally have the problem to choose an appropriate sample. Sometimes the sample is limited to a few dozen patients with a certain symptom or being treated with a medication that is in clinic testing. Here funding surely is a problem (who pays the study ;-)), but i think most are aware.

MDs do not write a dissertation, actually. None in the US, anyway (my wife's an MD, and we know more MDs than I can count).

 

 

[Green Baron]

Could be a German specialty ... anyway, the (German) medical dissertation isn't comparable to an engineering dissertation.

 

Edited November 19, 2016 by Green Baron

[PB666]

On 11/10/2016 at 1:26 PM, Darnok said:

I've said this for years... and they had to make study to confirm such simple thing...

http://phys.org/news/2016-11-criteria-funding-bad-science.html

Aside from being troll bait, if you wait long enough you will find a variety of opinions on the topic.

Being an assistant to the editor of 3 journals I would say a few things.

1. The quality of research of established science is improving. The quality of the statistics has greatly improved in the last 20 years. In some fields, from using nothing more than averages.

2. If you look at the number of rejected papers, this has greatly increased also, but not do to established science, but the science bomb that has come from India and China. Even these publications have increased quality.

3. Established science is doing more science, better science with less money (relative to GDP) than 25 years ago; however this will not go on indefinitely. The proportions of funding for small labs that is coming from NSF and NIH has declined greatly and the funding % of many NIH institutes might as well be zero, for small to medium sized labs. Pretty much all of the funding goes to large facilities with a dash of political impetus.

The nation that invest the most in science will be the leader in the next generation, and China is posing in that position (and they lack many of the myth based hangups that Americans have). Its not about the science it does, its about the techniques it can perform, the people it can train, the machines that it needs to build. A nation that can do these things for its science can also do it for its industries. When you look for instance at the biologic drugs, many of them can trace their origins or a comparable or like drug back to a university laboratory. And many of those invented by pharmaceuticals are simply copying the functionality of the laboratory produced chemical. You see pharmaceuticals touting the money spend on research but for cancer products they simply bought the rights from the university. Their research so-to-speak is promoting the drug for clinical trials and trying to get FDA approval.

 

[wumpus]

53 minutes ago, PB666 said:

You see pharmaceuticals touting the money spend on research but for cancer products they simply bought the rights from the university. Their research so-to-speak is promoting the drug for clinical trials and trying to get FDA approval.

To be honest, that is the expensive part.  Drugs that kill cancer in a petri dish are a dime a dozen (and every new science reporter breathlessly points out how their favorite isn't getting funding).  A bigger issue seems that those same companies can control publication (and the fact that null data wouldn't be published anyway only compounds the problem).  It simply compounds the issues of "p fishing".

I'm somewhat surprised about the "bad math" you are talking about.  ~25 years ago, my college roommate failed out of psych grad school due to poor math abilities (greatly annoying his EE PhD father).  I suppose that had he survived the class, he might 'revert to the mean' when writing papers.  Or maybe that generation (which was force to learn math) was what improved things so much.

I have to wonder if these studies are only possible now that the math used is good enough to test them.  I've seen similar complaints about the media, claiming that in the "good old days" reporting was serious and accurate (just don't look at H.L.Mencken's descriptions of how it was made in the "good old days").  The issue is that often you can get citizen-recorded video on youtube to check "the media"; in the "good old days" you took what they gave you (although everybody knew that "60 minutes" was incredibly biased against whoever they were smearing).

[KSK]

2 hours ago, PB666 said:

And many of those invented by pharmaceuticals are simply copying the functionality of the laboratory produced chemical. You see pharmaceuticals touting the money spend on research but for cancer products they simply bought the rights from the university. Their research so-to-speak is promoting the drug for clinical trials and trying to get FDA approval.

That's a bit of an over-generalisation. I'm not quite sure what you mean by 'copying the functionality' in this context, so please bear with me if I don't get this quite right.

If you're talking about finding a different chemical that can interact with the same biological target as the laboratory produced chemical, then that itself is not a trivial task and basically requires it's own research program. Knowing what target you're trying to hit is an important first step, no question about it, but it's only the first step of many and with modern molecular biology techniques, one of the more straightforward ones.

Target identification, compound library screening, hit to lead work and lead optimisation are well within the means of many university departments. Getting preclinical candidate compounds out of a university program is relatively common (which is not to disparage them in any way). Ideally, they'll be packaged up with as much preclinical data as the university(s) can put together before they're handed off to industry. Even then, there can be an awful long way to go (and still with a high drop-out rate) before that preclinical candidate makes it into clinical trials, let alone into patients. A pharma company licensing in a compound from a university program will still have a lot of development work to do, including formulation and scale up - which isn't a trivial task either - chemistry that works on the lab bench may be completely unsuitable for scaling up to commercial levels. Then, as @wumpus mentioned, the clinical trials and regulatory approval steps are the expensive part.

Regarding @wumpus's point about companies controlling publication, I'm optimistic that we're going in the right direction as regards publishing clinical trial data, negative results and all. See this article in Ars Technica for example. Regarding scientific results in general, standard contract terms for university-industry collaborations is to delay publication only for long enough to secure patent protection (if appropriate) and then publish the results. Any company that can't agree to those terms is going to find it hard to partner up with an academic institution. Publishing null data is probably still an issue but no more so than in scientific literature in general.

[PB666]

4 hours ago, wumpus said:

To be honest, that is the expensive part.  Drugs that kill cancer in a petri dish are a dime a dozen (and every new science reporter breathlessly points out how their favorite isn't getting funding).  A bigger issue seems that those same companies can control publication (and the fact that null data wouldn't be published anyway only compounds the problem).  It simply compounds the issues of "p fishing".

I'm somewhat surprised about the "bad math" you are talking about.  ~25 years ago, my college roommate failed out of psych grad school due to poor math abilities (greatly annoying his EE PhD father).  I suppose that had he survived the class, he might 'revert to the mean' when writing papers.  Or maybe that generation (which was force to learn math) was what improved things so much.

I have to wonder if these studies are only possible now that the math used is good enough to test them.  I've seen similar complaints about the media, claiming that in the "good old days" reporting was serious and accurate (just don't look at H.L.Mencken's descriptions of how it was made in the "good old days").  The issue is that often you can get citizen-recorded video on youtube to check "the media"; in the "good old days" you took what they gave you (although everybody knew that "60 minutes" was incredibly biased against whoever they were smearing).

The biologics are considerally more difficult to make than a simple antibiotic discovered in a dish. Often the antibodies start out in one species and end up being humanized (meaning information in one protein is replaced, bit by bit by information in another). There is a considerable amount of know-how and experience that goes into making some of these.

First you start with an antigen, It can be a whole protein or a segment of it. But thats actually not where it starts, because published X-ray crystallographic data has probably revealed a binding site or an enzymatic site you are interested in. So the first stage is to immunize into a model species such as Balb/C and generate an immune response. From this point you need to test for specificity and you need to find an assay that will give you the results you want. In the old days once you had specificity you would then create a hybridoma and then use multi-tier limiting dilution to screen the hybridoma lines for the specificities that are most similar to what you want (there is a whole process in that). We don't yet have a biologic, just a cell that produces something. Next the cells are expanded and injected in the peritoneal cavity of mice. The mice are then drained and you have an amount of antibody that you can used for testing in experimental animals. You will be doing this probably on 10 to 50 antibodies. Thats on the order of 50 mice per antibody. Meanwhile your cells are frozen into liquid nitrogen. The fast technique is gone now, pristane priming of mice has been pretty much ban by NIH for any institute that uses NIH money. The antibodies that make it to publication will be given a name like AMT-150 or B6H2F1 (limitiing dilution nomenclature). Then it goes through several rounds of publication and testing in different laboratories. Once the antibody has been tested the human on-variable parts of heavy chain and light chain will be genetically engineered into the mouse variable regions (1 or both chains, depends how the antibody works). This is then expressed in an E. Coli expression system. This whole process again is not one step, frequently does not work the first time, many PCR primers need to be made, remade . . . . .Then from the E.coli expression system you first have to test that it expresses, then you probably will have to move it to a system that optimizes folding and you can begin the first rounding of retesting. About this point big pharm is sniffing at your door, they might purchase here.  In may have to be moved to a eucaryotic expressions system, and still further retested. Next you have to absolutely purify if from all microbial contaminants. The Hep B vaccine that was made about 15 years ago had a small amount of contaminant that caused severe autoimmmune disease in some of the recipients. Once its purified you can retest it in a model in-vivo system. Then you have to begin testing in human volunteers  for tolerance, off to the prison system you go, but before you do that you have to establish its safe in animals. Next you can start looking, prolly in some country with relaxed protocols whether the drug is effective in humans. In very fatal diseases they may allow you test immediately. In one such test a drug was given and 6 of 12 recipients promptly died. 

And then the pharmaceutical companies just buy you out, the pay your university most of the price and you prolly get a new lab or a raise. Start to finish, around 15 years, the companies can do more quickly but they often don't publish what they have done, and some of the drugs have been pulled, so . . . . . After the wheel has been invented several times along of feild then big pharm starts making their own product de-novo, but in many cases we hardly know what it is that they produced. Etanercept is an example. https://en.wikipedia.org/wiki/Etanercept

 

 

[PB666]

2 hours ago, KSK said:

That's a bit of an over-generalisation. I'm not quite sure what you mean by 'copying the functionality' in this context, so please bear with me if I don't get this quite right.

If you're talking about finding a different chemical that can interact with the same biological target as the laboratory produced chemical, then that itself is not a trivial task and basically requires it's own research program. Knowing what target you're trying to hit is an important first step, no question about it, but it's only the first step of many and with modern molecular biology techniques, one of the more straightforward ones.

Target identification, compound library screening, hit to lead work and lead optimisation are well within the means of many university departments. Getting preclinical candidate compounds out of a university program is relatively common (which is not to disparage them in any way). Ideally, they'll be packaged up with as much preclinical data as the university(s) can put together before they're handed off to industry. Even then, there can be an awful long way to go (and still with a high drop-out rate) before that preclinical candidate makes it into clinical trials, let alone into patients. A pharma company licensing in a compound from a university program will still have a lot of development work to do, including formulation and scale up - which isn't a trivial task either - chemistry that works on the lab bench may be completely unsuitable for scaling up to commercial levels. Then, as @wumpus mentioned, the clinical trials and regulatory approval steps are the expensive part.

Regarding @wumpus's point about companies controlling publication, I'm optimistic that we're going in the right direction as regards publishing clinical trial data, negative results and all. See this article in Ars Technica for example. Regarding scientific results in general, standard contract terms for university-industry collaborations is to delay publication only for long enough to secure patent protection (if appropriate) and then publish the results. Any company that can't agree to those terms is going to find it hard to partner up with an academic institution. Publishing null data is probably still an issue but no more so than in scientific literature in general.

In the field of biologics in means for instance taking the specificity of an antibody, often humanized and optimizing it for more effect or less adverse effects.

It could mean, screening on a phage display library binding proteins, and then attaching that to polyglycine and the end of the antibody.

It could mean taking a rough domain and building a better peptide that has possible strain specific MHC class II antigen binding sites and can generate more antibody of the desired specificity.

Recloning. Since they are non-NIH funded and can operate anywhere they are free to use techniques essentially banned in universities. They generate and test more clones, maybe ten times as many sequence-motif specificities and select 3 or 4, versus 1. Next that can test them on different diseases, 1 may work better for one disease versus another. Of course they have a whole set up for humanization. YOu can make IgG of anyclass, or IgA or IgD depending on what you need. 

These fall under the category of 'likes', and frequently we have little information about what they are from a primary structural point of view. Etanercept is a like. Its an antibody, but not one you will ever find in nature.

Then there are a large number of hormones, cytokines, lymphokines, cognate receptors ect, that are sequence-farmed for their ability to generate decoy drugs that essentially titrate the level of a cross-talk agent in serum or in the environment of the pathology (site of cancer or inflammation). What they essentially do is take a sequence of a protein, maybe 10 to 30 residues, patent it, say any further use of it in is copyrighted.

 

"

The prototypic fusion protein was first synthesized and shown to be highly active and unusually stable as a modality for blockade of TNF in vivo in the early 1990s by Bruce A. Beutler, an academic researcher then at the University of Texas Southwestern Medical Center at Dallas, and his colleagues.^[2][3][4] These investigators also patented the protein,^[5] selling all rights to its use to Immunex, a biotechnology company that was acquired by Amgen in 2002.^[6]

"

 

 

[KSK]

15 hours ago, PB666 said:

Recloning. Since they are non-NIH funded and can operate anywhere they are free to use techniques essentially banned in universities. They generate and test more clones, maybe ten times as many sequence-motif specificities and select 3 or 4, versus 1. Next that can test them on different diseases, 1 may work better for one disease versus another. Of course they have a whole set up for humanization. YOu can make IgG of anyclass, or IgA or IgD depending on what you need. 

We're getting a bit off-topic so we should maybe continue this discussion on a separate thread but I'm curious to know more about these banned techniques. Humanization isn't banned so far as I'm aware and recloning is a standard technique - unless we're talking about different things here?

15 hours ago, PB666 said:

These fall under the category of 'likes', and frequently we have little information about what they are from a primary structural point of view. Etanercept is a like. Its an antibody, but not one you will ever find in nature.

Well no - it's an artificially engineered fusion protein so we won't find it in nature, or at least we're extremely unlikely to. We have plenty of information about it though, the primary sequence is published here (and that database is linked to by other sources so it seems to be legit). The European Medicines Agency seemed to be satisfied with the structural information provided as well.

15 hours ago, PB666 said:

Then there are a large number of hormones, cytokines, lymphokines, cognate receptors ect, that are sequence-farmed for their ability to generate decoy drugs that essentially titrate the level of a cross-talk agent in serum or in the environment of the pathology (site of cancer or inflammation). What they essentially do is take a sequence of a protein, maybe 10 to 30 residues, patent it, say any further use of it in is copyrighted.

Ouch. Personal bugbear - and not directed at you personally, since I see this a lot on the internet - but please don't mix up copyright and patents. Yes, they're both forms of IP but that's about it. They're very different legal frameworks intended to protect very different things. And speaking as somebody who works with patents for a living, I'd be very surprised if it's that simple. I very much doubt that they could get a blanket patent to a 10-30 residue fragment of a larger protein for example.

Edited November 21, 2016 by KSK

[wumpus]

3 hours ago, KSK said:

Ouch. Personal bugbear - and not directed at you personally, since I see this a lot on the internet - but please don't mix up copyright and patents. Yes, they're both forms of IP but that's about it. They're very different legal frameworks intended to protect very different things. And speaking as somebody who works with patents for a living, I'd be very surprised if it's that simple. I very much doubt that they could get a blanket patent to a 10-30 residue fragment of a larger protein for example.

https://www.ncbi.nlm.nih.gov/pubmed/11875409  ["How to publish DNA sequences with copyright protection."] [it's behind a paywall so I didn't read it, but if the NIH is publishing such via the Nature publishing group, that should be scary enough.]

While patenting and copyright may be designed to protect two different things, it looks like copyrighted genes have already happened.  This is a pretty big thing since there aren't any "copyright examiners", you just pay your $35 and get your copyright (possibly slightly more if you are storing large amounts of data).  And as you mentioned, they are different frameworks designed  to protect different systems, so all the exceptions to copyright law are unlikely to be useful for those using genes.

Note: while there were plenty of google hits on how to copyright genes, there was little to say if they held up in court.

[KSK]

Interesting - and alarming. Have to see if I can get hold of that because at face value it doesn't make much sense. I mean sure, I can write out a nucleotide or amino acid sequence longhand and it would qualify as a literary work under copyright law but unless making the actual chunk of DNA or protein counts as 'copying' that sequence then I don't see the point.

You're right though about exceptions not being useful. More worrying, the bar for originality in copyright law is quite low (you just need to prove that your work wasn't copied off anybody else), whereas the bar for novelty in patent law is fairly high. So an 'original' gene sequence (in copyright terms) doesn't necessarily mean that sequence is at all new.

Damn.

Mind you - the abstract for that article is a bit bizarre:

"I propose here an approach to encoding that will allow genomics companies to make their sequences available to the public while retaining some intellectual property (IP) protection. Using this approach, such companies would transform the DNA sequences in their databases into music files (e.g., MP3 format)."

[Listy]

The NIH didn't publish that article about copyrighting DNA - It's just hosted on pubMed, which is the NIH / National Library of Medicine's gigantic database of almost everything ever published in biomedical journals worldwide since the dawn of time - even silly, whimsical letters to the editor, like this particular piece was.

The letter is 15 years old & mostly a tongue in cheek commentary about mp3 file sharing services like Napster etc. It was written at a time when a DNA sequence potentially had some monetary value, because sequencing was hard, and slow and expensive. The author noted that natural sequences of DNA cannot be easily patented and cannot be copyrighted (a 'natural' DNA sequence is not a work of authorship) and so whimsically (the author was quite clear about that) proposed that by converting DNA sequences into mp3 files, the owner of a DNA sequence could charge a fee for other users for the right to 'listen' (decode) to that DNA sequence, and thus get rewarded for going to the effort of sequencing the DNA in the first place. Today, DNA sequencing is fairly cheap & quick, so trying to own or protect a natural DNA sequence like this today is pointless.

 

[Jonfliesgoats]

John Oliver did a bit about this.  

[KSK]

8 hours ago, Listy said:

The NIH didn't publish that article about copyrighting DNA - It's just hosted on pubMed, which is the NIH / National Library of Medicine's gigantic database of almost everything ever published in biomedical journals worldwide since the dawn of time - even silly, whimsical letters to the editor, like this particular piece was.

The letter is 15 years old & mostly a tongue in cheek commentary about mp3 file sharing services like Napster etc. It was written at a time when a DNA sequence potentially had some monetary value, because sequencing was hard, and slow and expensive. The author noted that natural sequences of DNA cannot be easily patented and cannot be copyrighted (a 'natural' DNA sequence is not a work of authorship) and so whimsically (the author was quite clear about that) proposed that by converting DNA sequences into mp3 files, the owner of a DNA sequence could charge a fee for other users for the right to 'listen' (decode) to that DNA sequence, and thus get rewarded for going to the effort of sequencing the DNA in the first place. Today, DNA sequencing is fairly cheap & quick, so trying to own or protect a natural DNA sequence like this today is pointless.

I wouldn't say its pointless if you're a company. Patenting natural DNA sequences is extremely difficult these days but if you could do an end-run around patent law and copyright those sequences instead, I could see a lot of commercial interest in that. Of course commercial interest isn't necessarily public interest but that's a whole other can of political worms that it's probably best not to open on this forum.

I've done a bit more digging on the internet and found various more-or-less convincing legal theories either for or against DNA copyrights. Ultimately, I think the idea will fail for policy reasons, particularly at the USPTO. After the recent high profile cases ruling against gene patents, I just can't see them reopening the whole mess by letting DNA copyright sneak past them.

[wumpus]

On 11/23/2016 at 4:00 AM, KSK said:

I've done a bit more digging on the internet and found various more-or-less convincing legal theories either for or against DNA copyrights. Ultimately, I think the idea will fail for policy reasons, particularly at the USPTO. After the recent high profile cases ruling against gene patents, I just can't see them reopening the whole mess by letting DNA copyright sneak past them.

Policy decisions by turf battle.  How all the critical decisions are made (then again, do you want Congressmen making arcane details about patents?  Reality and public perception have almost no correlation).

It's nice to know that paper wasn't terribly serious (and presumably the guy who wrote it new either that MP3s are lossy, and/or that they aren't reversible [enough for genes]).  You obviously can register such a copyright, enforcing it is another battle.