Live data from Hacker News

Scientists rename human genes to stop MS Excel from misreading them as dates

theverge.com

441–450 of 518 posts

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#441
post #410

Earlier quoted context omitted.

Since when is the universal way to identify a gene its locus? The locus is just the position of a given gene in a particular chromosome. The function of a gene is not exclusively tied to chromosomal location. If I take the sequence for SHH and move it to a different location in the genome under the appropriate promoters, it will still express sonic hedgehog. If I go and shove some other gene sequence into where SHH i…

do you know why HH is named such ? >>the function of a gene is not exclusively tied to chromosomal location "Since when is the universal way to identify a gene its locus" since molecular genetics...and gene mapping. this whole thing started with the idea that gene lables are arbitrary meaningless tags by extension not a problem to change labels to stop a spreadsheet program from mangling data. so now that we both kno…

You may note I used the word exclusively. TADs are hardly the end all be all of gene function. Epigenetics has more to do with gene regulation than function, and TADs aren't remotely close to the dominant force in epigenetics either.

Gene mapping through genetic linkage maps hasnt really been a major part of identifying and classifying genes for a while - not since sequencing became accessible. You could consider a genome sequence the ultimate physical (not linkage) map, but that's a couple decades out of fashion.

Edit: and im still not sure whether you're defending or attacking the naming of the hh family of genes, but yes, I'm intimately familiar with the hedgehog signaling pathway, and just because the hh loss of function makes fruit fly embryos look spiky doesn't mean that the name is justified or meaningful. Amd a basic problem in ontology is figuring out what that unifying theme should be to name all genes around, its not as easy as in chemistry. But until that's solved, the world loses nothing by changing shh to something less embarrassing so patients don't have to deal with being told their cancer has driver mutations in a pathway involving sonic the hedgehog.

Is your background in biology/molecular biology? I'm happy to share some links to resources if you'd like to read up on the state of the art now, including on gene ontology (GO). GO is an attempt to unify how genes are represented to give more meaning to the representations. I have my doubts about how successful these initiatives will be, but they're still really fascinating stuff. The field has come a long way since Morgan and Sturtevant developed that first linkage map back in 1913. And in epigenetics, there's so much more than TADs. There's the usual methylation/history acetylation etc, but also things like alternative splicing (perhaps not strictly epigenetics) that significantly change what genes do and what sequence even gets translated from the same locus. There's a lot of amazing stuff in happy to share some resources that I think are a good place to get caught up on the state of the art.

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#442
post #340

Earlier quoted context omitted.

That doesn't solve the analysis problem downstream when a non-expert is getting started, doesn't know about Excel's anti-features, and starts doing analysis using the only tool that the world has ever told them is acceptable for tabular data. Ideally excel would change, but since we know it won't, and we want to work with lots of people with minimal problems, we must adapt.

How are these non-experts getting started? If they are in academia, academia should start teaching these practical things too. (Yes, I know the problem is that many old school bigwig researchers are doing even worse things.)

There are people of all sorts: amateurs that want to play with the data but will not make study of it their primary field, trainees that will make use of tabular data extensively in their career, and then experts in more rarefied fields (e.g. immunology, clinicians, etc.)

If I can get a trained immunologist looking at my data, I'd much rather have 5 more minutes of their analytical skills than teaching them about common data exchange pitfalls.

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#443
post #119

Earlier quoted context omitted.

I don't even know why anyone still uses this notation in today's connected world. 2020-08-06 is ideal. Russian standard (06.08.2020) is uglier (digits are not sorted by significance) but still unambiguous. But that slash horror? Why would anyone continue inflicting this on themselves and their peers!?

>2020-08-06 I've just switched to 20200806. Let it be an int and be done with it.

greatideainthesameveinivedecidedtoforgospacesandpunctuationaswellbecauseitsavesspaceaswellaswearandtearonthekeyboard

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#444

It's honestly amazing that Excel hasn't fixed this issue. It's pisses off an enormous number of users especially in basically any non-US country (even if 01/02 is a date, it isn't the second of January in most of the world...)

Why hasn’t the field standardized around another tool? Is the list of excel plugins that something libreoffice couldn’t replace it, where they could just make add the options they need? My guess is the field is mostly lacking people interested in working on tooling like this?

It mostly has; most serious work is done in R and python, and there are a lot of supporting tools, documentation, and useful packages for both languages. It's not like the actual statistical analysis in your average paper was actually done in excel.

However, Excel remains a nice tool for "I'll just look at this CSV with the final results from this analysis, sort it by correlation, and see if any of the usual suspects are up top". And if the next step is "yeah, that looks fine - I'll just copy the top 100 genes into this convenient GUI pathway analysis tool", you're suddenly exposed to whatever Excel did to your data.

And as for "why not libreoffice", most researchers I personally run into are strong molecular biologists who've learned a subset of R for their uses; they're not really likely to go out and find libreoffice on their own. Besides, the writing process for papers includes sending drafts and spreadsheets to doctors and pathologists and editors, who are probably on hospital computers with a short whitelist of programs ... and I don't really want to debug subtle compatibility issues in the sort of garbage fire those documents can turn into.

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#445
post #384

Earlier quoted context omitted.

Excel is stuck. It can't change this behavior if it cares about backwards compatibility. There are uncountably many recorded macros and all sorts of scripts and data tools everywhere, that implicitly and inadvertently depend on behavior like this. Like Javascript's warts, you have to declare these as features and live with them. Of course Microsoft has thought about this, it's silly to think they never have and we're…

> It can't change this behavior They could easily make it an optional behaviour.

Exactly, add a checkbox in the settings menu and the problem is solved for everyone.

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#446

Earlier quoted context omitted.

> "01/02" does not translate to Jan 2nd in most of the world, because DMY is much more popular than MDY. This doesn't follow. There's no year in 01/02. MD is more popular than DM.

In which country do people use DMY, but also MD? This just sounds wrong, and my (admittedly limited) sampling shows that when DMY is abbreviated it becomes DM. 01/02 is still the 1st of February everywhere I went except in the US.

Who said someone used DMY combined with MD? DMY/MDY is not an exhaustive list of date formats, you know. It is not even an exhaustive list of "incredibly common date formats".

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#447

Earlier quoted context omitted.

From my point of view Excel hits a sweet spot between 'very simple tasks' and 'very complex tasks'.

The sweet spot being "too complicated for simple tasks" and "not sophisticated enough for complex tasks"?

>"too complicated for simple tasks"

How can Excel possibly be too complicated for simple tasks? It is pretty much as straightforward as it goes when it comes to grid-file viewing and editing. You can show it to anyone from a high-schooler to a 60 year old (with minimal experience on computers) colleagues and they will figure it out rather easily, good luck teaching Python/Pandas to the latter.

>"not sophisticated enough for complex tasks"

Not sure how that works either really. Between formulas and VBA macros, people have and are making tools complex enough they have no business to be an Excel, and yet they are even if it isn't the best tool for it. Once you go past that point, Excel isn't even in the conversation nor does it pretend to be able to.

It has issues, and people playing or working with complex (or simple) data would be better served to learn programmatical tools, but until they do Excel will serve them well as long as they stay wary of basic quirks.

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#448
post #447

Earlier quoted context omitted.

The sweet spot being "too complicated for simple tasks" and "not sophisticated enough for complex tasks"?

>"too complicated for simple tasks" How can Excel possibly be too complicated for simple tasks? It is pretty much as straightforward as it goes when it comes to grid-file viewing and editing. You can show it to anyone from a high-schooler to a 60 year old (with minimal experience on computers) colleagues and they will figure it out rather easily, good luck teaching Python/Pandas to the latter. >"not sophisticated eno…

You could presumably come up with similar arguments for why Word (or equivalents) isn't "too complicated for simple tasks" and "not sophisticated enough for complex tasks", but, from experience, I'd argue that's also incorrect.

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#449

Earlier quoted context omitted.

imagine if they worked with the industry to build in a set of relatively agreed upon defaults, like turning autoformat off for dates that matched a critera. then geneticists could activate the geneticist profile and excel works as they'd expect. if they encountered a file that broke, they could turn that profile off. that's a fairly sophisticated solution, but it's just one of many potential approaches to fixing the…

It would be way less effort to have the same exact program under a completely new name. MS Excel MS (names are harder than I thought / remembered) Matrix In the back-end they'd use the same sanity and share 99% of the code, but Excel would add in the cruft / warts compatibility quirks.

Less effort and they can charge twice for the same product, genius!

Re: Scientists rename human genes to stop MS Excel from misreading them as dates

#450

Earlier quoted context omitted.

Excel is from an era when programs still catered to power users. Tools were made to have learning curves, ideally not particularly steep curves, but curves nevertheless. It wasn't expected that users would hit the app running, intuiting everything there was to know about the program in their first minute of using it. The result is a rich deep program that users can grow into, rather than a shallow trivial program tha…

Auto-formatting seems like more of a noob feature than a power user feature. Lacking control over when it happens is definitely not a power user feature.

Treating SEPT1 as a date is behavior that will be correct for the vast majority of users, be they noob or experienced, and only incorrect for a very tiny minority of users who are doing things related to genetics. This sort of auto-completion feature is orthogonal to the noob/power user axis, except insofar as the user is expected to know how to circumvent this behavior if/when they need to.
Post reply on HN