In the last 36 hours we have 21 new questions (~14 / day). With this tempo we have really nice chances.

I am afraid what we need are new users though, not just new posts from already active users. That's what Robert's meta post said.

> Fair warning: Simply having the same people ask more questions is not sustainable and not a solution for giving the appearance of organic growth. You need more users.

Is the site in risk of being closed? ( Poor Bioinformatics.SE ;( )

Upvote this post on reddit please, so we can have more people to join.

@kenorb I thought we were already covered by:

but ok.

People at work are really excited about this new forum, but excitement is dampened somewhat by not being able to sign up...

@Kusalananda They can if they go through area51, I think

Alternatively, send them invites from the form on the right hand side of the main site.

Thanks @terdon!

Hey, thank you!

@Kusalananda Note StackExchange is not a forum it's a network of Q&A sites.

Just want to get the message across.

@HenryWHHackv2.0 I know. Sorry. Not a forum.

Ok

Though, a place where one could ask questions and have them answered is technically a forum...

But I will not use that word about the SE site if it's wrong.

Related, I suppose: meta.stackexchange.com/questions/92107/…

yeah, SE folks have a bee in their bonnet about being referred to as a forum.

Largely because it implies discussion.

Also because many think the word only refers to online forums

I was just using the word because I couldn't think of any other word. "Q&A site" is ok with me.

I think what the site is missing at the moment is ironically softball questions

So far, most/all? active members seem to be relatively experienced in the field, and consequently the questions are mostly quite hard, and research-level

@terdon we do have quite a lot of new users (or at least some new users got active).

What the site needs is a few questions on undergrad/masters level that are tricky to answer on one’s own but can be answered authoritatively by “working bioinformaticians”

@KamilSJaron Yes, that's what I'm pinning my hopes on.

@KamilSJaron Thanks for the link to REPET, didn’t know that one yet

@KonradRudolph you mean like "How to plot histogram of lengths of sequences in fasta?"

@KamilSJaron Think simpler.

Things like "what is a PWM?" I guess. Just basic level stuff.

@KonradRudolph :-) perks of non-model organism research...

@terdon Yeah, maybe, although this raises the problem at what point questions are too basic for the site. PWM for instance has its own Wikipedia article

And yes, some programming questions would be nice.

@terdon How to count number of sequences in multifasta?

@KonradRudolph True.

@KamilSJaron Heh, I guess although we probably won't want things quite as simple as that if/when we graduate.

By the way, I would think that basic *nix questions that would be faced by a bioinformatician should also be on topic (like how can I count the no of sequences in a fasta file).

Even "How can I install $bioinfoToolX on $*nixSystemY" etc

Or Windows, I guess, if people are actually using that.

@terdon it is not that long time ago when I graduated... These too simple questions could be easily translated into questions with non-trivial answers by adding "What is the most efficient ..." at the beginning.

True.

Although I doubt you'll be able to beat grep -c '^/>/' file.fa for efficiency.

@terdon there will be probably more efficient C solution, which creates a nice opportunity for more nice answers...

@KamilSJaron I honestly doubt it, actually. Not unless you use an index file. If not, C will do pretty much what that grep does.

@terdon Maybe. There is only one way how to find out...

Heh. let me know if you do code it :)

It would be simple in C, but it also leave lots of room to screw it up. It's IMHO better to use a basic Unix tool for things like that.

@Kusalananda I definitely agree that using standard Unix tools is better / wiser / more reliable solution. I was just not entirely sure if it is the fastest solution...

@terdon probably fgrep -c '>' foo.fa …

Although no, that’s probably less efficient because your regex can stop reading after the first character in a line

@KonradRudolph That's just grep -Fc, yes.

Ah yes, the ^ is essential.

Otherwise it's slow as hell.

-F

-F, true

Bloody lurking sysadmins!

:P

:-) It would also fail if a header line contained multiple >

@Kusalananda Not with the ^

No, sorry. It wouldn't fail. The number of lines would still be the same.

... but if the sequence contained >... Can it?

No

Kinda the point of the FASTA format.

What about FastQ?

The sequence can only contain A-Za-z and -

Ok.

@Kusalananda That's different. The headers aren't defined by > but instead by @ and yes, the @ can also be found in the quality line.

Ok. I'm not a user of those formats myself, only the occasional filter writer.

You have to do things like echo $(( $(wc -l < file) /4 )) for fastq.

Or awk 'NR%4==0{n++} END{print n}

Much more of a pain to count.

Or just END { print NR / 4 }

Actually looks like an easier format to parse than Fasta, since it's always 4 lines per sequence...

Well defined records.

@Kusalananda Duh! Of course.

@terdon That’ll fail with some older FASTQ files unfortunately, because the sequence (and quality) can actually span multiple lines.

And counting /^@/ isn’t enough either because as mentioned the quality line can start with @.

So yes, FASTQ really is a pain in the neck

It actually needs a small state machine to be parsed; in minimal readable Perl that’s going to be at least ~15 lines of code.

@KonradRudolph Ah, yes, I was wondering about that. I could see no reason why the format would be limited to short reads, but I assumed they'd enforce very long lines instead.

TO make everybody happy : bioinformatics.stackexchange.com/questions/361/…

@KonradRudolph what about grep "^[ATGCN]"?

@IakovDavydov Well theoretically a quality string in some Phred encodings could consist of nothing but these letters … also, don’t forget about U

(for RNA)

@KonradRudolph if it has letters, and read has multiple lines, it becomes impossible to decode the file deterministic ally, no?

@IakovDavydov Yes, because the sequence and the quality need to follow the same “shape”, i.e. if the sequence is broken into lines after 5, 10, and 5 characters, then so must the quality score.

Again, see my example Gist:

gist.github.com/klmr/81a2b86cd93c706d28611f40bdfe2702

Wait, in fact I’m not even sure about what I just said. But the lengths of the sequence and the quality must obviously be the same

So what the script does is record the sequence length, and read that many characters in the quality string. What this means, though, is that the FASTQ format is not a regular language.

(Though it’s context free)

Anyway, done for today, have a nice (UTC) evening