James Hawley @jrhawley@scholar.social

I'm a PhD candidate in Medical Biophysics at the University of Toronto studying epigenetics and cancer. Math and physics background taking a step into bioinformatics and evolution #introductions

This is a good article on #openscience and #opendata from Nature

https://www.nature.com/articles/d41586-019-03335-4

I recommend looking at the list at the end on papers you can read, discussions you can have, and long term goals you can set over 10 weeks for transitioning to more transparent and collaborative scientific practices

Off to the Fields Medal Public Symposium! It's always nice to see how mathematicians present their work to the public, especially brilliant ones

Often, papers Gelman discusses have some type of fallacy in them, so here's his handy guide of things to look out for:

https://statmodeling.stat.columbia.edu/2009/05/24/handy_statistic/

Here's an important quote that really highlights something present in my own field of #genomics and #epigenetics too:

> And that’s why the authors’ claim that fixing the errors “does not change the conclusion of the paper” is both ridiculous and all too true. It’s ridiculous because one of the key claims is entirely based on a statistically significant p-value that is no longer there. But the claim is true because the real “conclusion of the paper” doesn’t depend on any of its details

It contains an extensive list of a history of articles on reproducibility in psychology as well as a discussion of how these ideas play out in everyday scientific publishing, and what it means more psychology specifically and science more generally

Excellent article by Andrew Gelman

https://statmodeling.stat.columbia.edu/2016/09/21/what-has-happened-down-here-is-the-winds-have-changed/

I like that he's not going against Fiske, herself, but about the ideas that she has used in articles as used in judging articles as an editor

The US Census Bureau is already making use of these algorithms and has them available for people to look at on GitHub

https://github.com/uscensusbureau/census2020-das-e2e

I love this kind of research that puts individuals first, while maximizing the utility of their information for the public good

Her paper can be found here https://arxiv.org/abs/1909.12237

I look forward to reading this and seeing these algorithms in action

I attended a talk by Ruobin Gong (Rutgers University) about differential privacy and the 2020 US census bureau. This is super cool stuff that provides the foundation for a paradigm shift in how data is stored and analyzed in public.

By perturbing individual-level data with known distributions that maintain privacy of each individual against certain kinds of queries, privacy and transparency don't have to be at odds with one another

> From now on, all our research content will also be published in the ‘Article’ format; the shorter, ‘Letter’ format has been retired.

I couldn't be happier with this decision. Nature finally realized that their Letter format sucked, and authors were treating them like articles just without headers or any hope of being readable.

Making them all articles is definitely the way to go, I greatly appreciate this change

https://www.nature.com/articles/d41586-019-03167-2

There are some really interesting ideas in this paper about genetic databases and privacy. It's sort of like finding out what email addresses are in a database by spamming the log in and seeing what responses you get.

https://www.biorxiv.org/content/10.1101/798272v1

Followed by a brief discussion in this Twitter thread
https://twitter.com/DocEdge85/status/1186645459617513472

Did the same graphic designer do work for both JetBrains and the Francis Crick Institute? Logos are remarkably similar in style

Their final figure sums up their results quite nicely! Please share with your colleagues, if relevant

Here's a really good paper from @keegankorthauer@twitter.com on methods controlling false discovery rates.

https://doi.org/10.1186/s13059-019-1716-1

While they focus on #genomics, they give a good overview on what #fdr and #fwer controlling methods do and what they're for that's applicable to anyone in any field performing multiple simultaneous hypothesis tests