Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The actual paper is at http://www.nature.com/nm/journal/v22/n8/full/nm.4133.html

I was going to see if it could be torn apart but

1) Shen-Orr is a very thorough, un-full-of-shit researcher,

2) I like how they used CyTOF in a novel capacity here, and

3) depressed people do tend to get sick more often.

That said, some of the largest divergences between mice and primates are in the brain, so it would be wise to exercise caution when generalizing these findings to humans.

If you need the full paper and have misplaced your subscription credentials, I hope you would not resort to plugging the DOI into http://sci-hub.bz like so:

http://sci-hub.bz/10.1038/nm.4133

Because that would be wrong. Publishers need hot tubs in their 3rd bathrooms, just like Deans and coaches do.

The methods section is included in that version (papers without methods are like dehydrated water). Read it for yourself, don't listen to some blowhard on the Interwebs (i.e., me).



as a big proponent of the cytof, i'm not sure that these guys used it in a novel capacity or even used it in its area of greatest strength... based off of what i can see, they did some simple immunophenotyping with less than a dozen markers, which traditionally be more suited for flow cytometry-- which they also used to detect GFP in a different section.

i like that the cytof is getting more popular, but people need to understand that the logistical commitment and jankiness of the system means that it should really only be used for its greatest strength: simultaneous detection of 30 variables on a small scale basis.


nah its biggest strength is to use as IMC for tissue sections ;-)

but yes, reading more carefully, I suppose you're right. Then again, orthogonal validation is not a bad thing.


Curious, because I don't see it offered. Did they publish the statistical code and (anonymized) data?


The code for Citrus is on GitHub (see online methods). The data was deposited into Cytobank, so if it isn't already available, you could ask them to release it (Cytobank is sort of like GEO for flow/CyTOF data). It's pretty transparent.


ps. Mouse data doesn't need to be anonymized :-)


Mouse HIPAA wants to have words with you, apathy! What about mouse privacy?! Despicable, that's what it all is, despicable!!! :D ;) :)


So basically, "no". Citrus seems like the algo, not the glue code or total source, if you will. And the data isn't published freely. :(


Most of the libraries for analyzing CyTOF data are already open (from Fluidigm). Did you already check for their data on Cytobank?

Granted Nat Med is not as demanding re: reproducibility as say eLife or Bioinformatics, but they're not complete shits like Cell Press either. Ask the authors to open the Cytobank data if they haven't. I don't see why they wouldn't.


> un-full-of-shit researcher

Not a very strong statement ;)


I just spent a day hanging out with him. A very thoughtful, smart person who does innovative work.


Oh, you might be surprised


That depends on who wrote it, and other context.


It's like saying un-poisonous food.


Well, that depends on the dose, doesn't it?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: