Ashish, I've long been a fan of your work and am excited to see you share InpharmD here.
1. What are your thoughts about partnering or selling to pharma / biotech / R&D orgs? Is there a potential value prop?
2. How does this compete with or complement existing clinical informatics and medical librarian capacity at academic medical centers? Or are they not the target market because they already pay salaries for humans to do these tasks? How do above entities relate to what the PharmDs do?
Excellent questions and thank you for the kind words :-)
1. We tested this with Pfizer last year and found there was an opportunity to supplement existing med info teams that do the same thing.
But it’s tough to do two markets well at once, so we decided to focus on health systems for now.
We also find that in hospitals, everyone thinks they’re asking b unique questions, but they aren’t. We can be much cheaper vs their pharmacists and still make money. Pharma companies already have standard responses so it’s a totally different value prop.
2. We don’t really compete with clinical folks at hospitals, most will readily off load this to us, so they can spend more time on patient care. There are some that like to own this, and I totally get why, but we eventually win them over. As for medical librarians, they’re great for article requests but for complex clinical questions we think a clinical pharmacist is the right type of researcher.
Estimating a minimum required sample size is one of the most common questions asked by clinical or biomedical collaborators before embarking on a research project. This is especially true when ML is an option. This paper provides rules of thumb and a digestible amount of theory that could inform such conversations, and will surely become a popular reference.
Note intuition from traditional statistics does not universally apply to deep learning and/or extremely high-dimensional data. For example, deep neural networks with 1-4 orders of magnitude more parameters than training examples can still generalize well to unseen data.
They develop open-source platforms that are inspired by and seek to solve problems lived firsthand by collaborators (such as myself) and the greater scientific community.
The Broad has created an environment and provided resources to enable engineers to do great work. The DSP feels like a blend between tech company and academic research institution. The people I've interacted with are technically superb (and also nice).
If you are interested in contributing to the intersection of life science and compute, would strongly recommend you check them out.
This may be off topic if you consider above to not be sufficiently "blue collar", but biomedicine in general (from the entire spectrum of basic science all the way to care delivery) is in dire need of smart computational folks.
This article does NOT provide compelling evidence that fungi are the causative agent for AD. It merely reports correlative observations.
You may conclude that antifungal drugs could help patients with AD. You may also think that vaccination or some other mechanism of prevention could reduce the prevalence of AD. These conclusions are simply not supported by the data.
I am a biomedical data scientist, not a neuroimmunologist, so I asked a colleague in my MD/PhD program who has experience in the latter field. His thoughts:
"There are a number of concerning issues here:
1) They are using polyclonal antibodies. I don't think they addressed possibility of cross-reactivity. Comparing Alzheimer's disease vs control brain is like comparing apples and oranges. There can be very different inflammatory states between the two and yield different antigenic environments.
2) The possibility that neurofibrillary tangles or amyloid plaques are sticky and can non-specifically bind antibodies remains a possibility.
3) Only immunohistochemistry? Could have at least done some qPCR especially since they can grow these fungi for quantiation. The only other paper that observed fungus in CNS of AD patients is by the same group. The high possibility of artifact is has not been ruled out."
Have you considered / is there already a way to “import” an .ipynb into a DeepNote project?
Would be easier for users than to copy and paste code one cell at a time.