Which New Bioinformatics Related Tool Would You Appreciate The Most?

Question

[Biomonika (Noolean)]

I am about to spend next 6 months working on my thesis, part of which should be some new bioinformatics related tool. I would love to do something, what could be useful (and used) also for broader scientific community. All my ideas so far have been either too complex, too easy to do or one-use (pipeline, which cannot be used repeatedly). I can do desktop applications with proper GUI, as well as web interfaces and mobile applications.

Don't be limited by technology, think about applications:-) (I will learn whatever will be needed even if I miss piece of knowledge now) So, my question is:

Which new tool would help you? What do you miss? What already existing tool would you like to replace/improve?

Thanks a lot for suggestions.

[Pierre Lindenbaum]

related: http://www.biostars.org/post/show/494

[Mary]

Oh, yeah, I still completely want that tool I described there.

[Michael]

You see, this is place hopeless for a question like this, not a single original proposal yet.

Answer 1

[Michael]

I had earlier written a forum topic on choosing a thesis topic. As you can see, nobody else has contributed. That could be due to the reasons I have stated in the post, the low visibility of the forum compared to the questions or, more skeptically speaking, people on this forum are not inventive enough. ;)

So, let's give it some thought again about some neat tools (notwithstanding that some of these maybe already exist and are not totally inventive themselves):

• There are a lot of draft genome assemblies these days; develop a tool that automatically infers contradictions and problems (including a meta-genome approach to contamination in the sequences) in draft assemblies, based on all available data sources (paired end reads, fosmid libraries, EST sequences, conservation scores, and what not) and visualizes them in a good way.

• There are a lot of draft annotations of draft genomes; make a good software that automatically detects flaws and inconsistencies in gene predictions and genome annotations based on a large number of data resources (e.g. conservation, alignments, RNA-seq, expression data, etc. ) and visualizes them.

  Discuss

[Michael]

Oh and btw, if you want to come here to do your thesis, I might even supervise you :D

[Biomonika (Noolean)]

Hi! I especially like the first proposal, as I do not have experience with genome annotations, but I did work with Newbler and MIRA (for example) and I know how much their outputs differed :-) I am now thinking how I could use resources you mentioned (its not easy at all:-) and unfortunately my employer (and supervisor-to-be) is as a biologist not really experienced with these tools.

[Michael]

Yes, it might be a bit over the top. What is the specialty of your boss then? Also, couldn't you -totally innocently - ask him or her exactly this question, (just like sometimes you ask what people like as a birthday gift)?

[Biomonika (Noolean)]

Hi Michael! My boss expertise area coveres mostly transposable elements, cytogenetics and microsatellites. However, I asked everybody in our lab with following answer: http://biostars.org/post/show/50662/from-illumina-data-to-x-linked-and-y-linked-genes/

Answer 2

[Casey Bergman]

I would love to see a fully-fledged SAM-to-SOAP format converter to allow use of any short read mapper to be combined with SOAPsnp's consensus sequence module. See more on this topic here.

Answer 3

[ff.cc.cc]

I'm aware that a couple of tools offer a visual interface to R debugging:

i) revolutionanalytics (free for academic use) sells its own environment with some closed-source features

ii) StatET is an environment derived from a customization of the eclipse SDK, I suppose it is open source and feature

...but, for many reasons (...is there a thread about R environment ?) I don't love them, instead I feel very comfortable with RStudio. So I suggest to finally develop a full-featured visual debugger for R that fits well in this environment.

Other appreciated (just my opinion) R features regard interactive-graphics package for big data visualization.

Have nice job !

[Biomonika (Noolean)]

Thank you a lot for suggestion! I am R beginner, so I am unfortunately not familiar with these tools, used just command-line version:/ However, I still think that this is far more complex than for a one man show.

Answer 4

[lgbi]

We often need to blast hundreds of 16S sequences and write per sequence the top < x > results in a table. This way we have a quick overview of what species our bacterial community may consist of.

I personally do it with Biopython scripts or using standalone BLAST+, but I'm sure my wet-lab colleagues would prefer a GUI or webinterface.

It shouldn't be too difficult to create something like that for someone with programming experience.

[Michael]

hm, ViroBlast, NCBI Web-Blast? I think that is a bit too easy for a master's project.

[lgbi]

ViroBlast is only for viral sequences. NCBI Web-Blast gives results on a separate page for each input sequence: if you have a few 100s of them, you'll have to click a lot.

Too easy... maybe, but there are some challenges and six months is not a lot. Of course, you can do it in 2 weeks and throw it on the pile of existing bioinformatics software which is not made to be used by anyone but the creator :)

[Michael]

"ViroBlast is only for viral sequences", no. You can use that interface for anything if you download and install the code, it's just a web interface to blast. In addition I simply believe this task is not original, because it has been solve many times.

[Biomonika (Noolean)]

Thanks. Yes, it is easy (as Michael suggested). However, I think that set of simpler programs could be as useful as one complex. And we could also think about extending this project, adding new features.. Can you imagine what else could be done with this kind of data (some visualisation maybe)? :-)

[lgbi]

You could include taxonomic information based on the NCBI taxonomic databases to get species, genus, family, ... Visualisation of for example genus frequencies is often used in publications.

[Michael]

Like MEGAN? Are you thinking about building something like "WebMEGAN" with integrated BLAST runs?

Answer 5

[JC]

I really like the web user interface for several tools, such as NCBI-Blast, Ensembl and Galaxy.

[Biomonika (Noolean)]

So, for what tool would you like the new user interface to be created?

[JC]

I like Galaxy, really easy to use

[Michael]

Weren't we talking about something new? (New, as in "doesn't exist yet").