Time To Rehash: How and why I choose my sources

A year ago, I wrote a post (“Sources Named: Who I quote and why“) explaining how and why I choose my sources.  I want to know something is an original source, or a retelling or re-reporting.  Sometimes, the source ends up being fifth hand, with very different wording or meaning than the original.  It ends up being a game of telegraph rather than telephone.

I also said I prefer scientific sources or university publication over news sources.  Commercial news likes to embellish, either to sell fear or false hope rather than print boring facts.  When was the last time you saw “study reconfirms theory”?  An item from the past week shows why I do this.



Taiwan News is “journalism” on par with The Express (UK) or “weekly world news”.  I don’t believe anything TN posts, even when it goes back to the original source and its hard to get wrong (e.g. earthquake reports).  So when they published an item last week (9/23) about a “cure for diabetes”, my immediate response was, “Yeah, right.”

Taiwanese scientists discover potential cure for diabetes

Taiwanese scientists have discovered a key mechanism that causes diabetes and developed a new drug treatment that could “fully reverse” the disease, according to reports.

Diabetes has been shown to be caused by the loss and function of Beta cells (β-cells) in pancreatic islets, which are regions of the pancreas that contain hormone-producing cells. Yang Wen-chin (楊文欽) and his team at Academia Sinica’s Agricultural Biotechnology Research Center (ABRC) have discovered through experiments on mice that the protein-coding gene, Pdia4 (Protein Disulfide Isomerase Family A Member 4), is responsible for the destruction of β-cells and that inhibiting this gene can prevent and even reverse the loss of such cells.

The so-called “news” mentions a couple of names, but zero sources.  I’ve been waiting for a better report, and actively seeking scientific papers to see how wrong Taiwan News was.

On Saturday (9/25), the Taipei Times showed why sources matter.  Not only do they more accurately call this a treatment for diabetes, not a “cure”, they also cite the scientific publication where it was published.  Now I consider this a credible story.

Study finds diabetes-linked gene

Taiwanese researchers have identified a gene that they say might help doctors treat diabetes.

A study by a research team at Academia Sinica’s Agricultural Biotechnology Research Center showed that the expression of protein disulfide isomerase family A member 4 (PDIA4) is linked to diabetes.

The research was published in this month’s issue of EMBO Molecular Medicine, an open-source, peer-reviewed scientific journal based in Europe.

Details make a difference – who wrote it, where it’s published, and how to find it and then read it.  The details of the paper are waaayyy over my head (though I’m sure PZM understands it), but you can get the gist of it from the abstract (page 1 of the PDF, and below) and Paper Explained (page 19):

Pdia4 regulates β-cell pathogenesis in diabetes: molecular mechanism and targeted therapy


Loss of β-cell number and function is a hallmark of diabetes. β-cell preservation is emerging as a promising strategy to treat and reverse diabetes. Here, we first found that Pdia4 was primarily expressed in β-cells. This expression was up-regulated in β-cells and blood of mice in response to excess nutrients. Ablation of Pdia4 alleviated diabetes as shown by reduced islet destruction, blood glucose and HbA1c, reactive oxygen species (ROS), and increased insulin secretion in diabetic mice. Strikingly, this ablation alone or in combination with food reduction could fully reverse diabetes. Conversely, overexpression of Pdia4 had the opposite pathophysiological outcomes in the mice. In addition, Pdia4 positively regulated β-cell death, dysfunction, and ROS production. Mechanistic studies demonstrated that Pdia4 increased ROS content in β-cells via its action on the pathway of Ndufs3 and p22phox. Finally, we found that 2-b-D-glucopyranosyloxy1-hydroxytrideca 5,7,9,11-tetrayne(GHTT), a Pdia4 inhibitor, suppressed diabetic development in diabetic mice. These findings characterize Pdia4 as a crucial regulator of β-cell pathogenesis and diabetes, suggesting Pdia4 is a novel therapeutic and diagnostic target of diabetes.

I said nineteen months ago I would amend, update, and link to the 2020 post on the main page.  Oops.


  1. robert79 says

    “Taiwanese scientists have discovered…”

    I really, really, really, despise reporting like this, and a lot of major newspapers do it.
    – People did this stuff, they deserve to be named. (Yeah… they’re doing it “for science” their paychecks and promotions still depend on stuff like this!)
    – If it’s interesting and somewhat in my field of expertise, I may want to know more… not more newspaper fluff, but the actual research. And no, just because I try to read the scientific literature as well doesn’t mean I actually have run across the paper that spawned a press release that spawned the interest of some reporter, and *only* knowing the country and the subject makes it quite hard to find a paper.

    I suspect it has to do with how university PR people format press releases or something, but even then it makes *no* sense, since I’d think a university PR dept. would want to actually *name* their university…

    The weirdest bit is when the reporters then go to some ‘expert’, and do mention them by name/university… “Professor of university (unrelated to the study) said this is a remarkable result…”

  2. Bruce says

    Too many media places give their “science beat” to the applicant who lost on the competition for the fashion beat. That is, it’s rare for the job to be held by anyone with the inclination or background to do the job, let alone both. Editors too often reflect the trend that the most science-y possible item is one on ancient interstellar aliens.

    Many decades ago in North America, and just a few decades ago in most of the rest of the world, the main form of diabetes was what we now call type 1. We should note that diabetes is now two different diseases. Type 1 is not making insulin, which is what I think the study was trying to talk about.
    In countries with a Standard American Diet, there has even a huge rise in type 2 diabetes, connected to obesity, hypertension, fatty liver, etc. And while severe cases are often treated with insulin, the real problem is the body’s learned insensitivity to insulin, due to a long-standing overdose of insulin, due to eating changes, such as high carbohydrate/sugar foods and drinks, snacking between meals, fear of healthy fats, and so on.
    Decent journalism on this major issue would at least point out that this research focused on type 1, while most overweight people around the world are in danger from type 2, which is not associated with any problem in one’s pancreas in secreting insulin.
    Thanks for looking in to this story.

  3. naturalcynic says

    @ 2. Bruce:
    This study almost certainly is mostly concerned with the late-stage pathogenesis of Type 2. From the descriptions in the Intro. :
    Globally, 425 million people live with diabetes, which causes about
    5 million deaths annually. Diabetes is characterized by a failure of
    functional b-cells to adapt insulin secretion to compensate for
    increasing insulin resistance, driving diabetes development …

    Insulin resistance is the original problem with Type 2 – the demographics have to mean Type 2. In humans and experimental animals, later stages of Type 2 show an eventual significant decrease in beta-cell competence in secreting insulin and develop into slow beta-cell death. Early-stage Type 2s show hyperinsulinemia as a response to loss of insulin sensitivity, but as the years progress [sometimes decades in humans] the beta cells eventually die out and the patient needs to go on insulin therapy – if the subjects live that long. Beta cells show a slow pathological process rather than the relatively rapid pathology seen in Type1s. A few years ago this was described as “they just get worn out”. This paper seems to describe this slow pathogenesis. I wonder whether the pathologies of the two[+] types may be more similar than originally thought.
    Intransitive gives the link, so I glanced at the article [20 or so pages] with lots of figures. I’ll have to take my time with it …