101 questions with a bioinformatician #34: Katie Pollard

This post is part of a series that interviews some notable bioinformaticians to get their views on various aspects of bioinformatics research. Hopefully these answers will prove useful to others in the field, especially to those who are just starting their bioinformatics careers.

Katie Pollard is a Senior Investigator at Gladstone Institutes and a Professor in the Department of Epidemiology and Biostatistics at UC San Francisco. She is also a Faculty supervisor of a bioinformatics core that provides collaborative support for high-throughput biology across the UCSF campus.

Katie's work involves the development of statistical and computational methods for the analysis of large genomic datasets, with a particular interest in genome evolution and identifying sequences that differ significantly between or within species. Her work on the chimpanzee genome has led to lots of coverage by mainstream media, and if you want to know more about this topic, you should definitely watch the What makes us human? talk that she gave at the California Academy of Sciences (video is online here).

You can find out more about Katie by visiting her lab's website. And now, on to the 101 questions...



001. What's something that you enjoy about current bioinformatics research?

Growth in new sources of data, such as from citizen science and electronic medical records, as well as emerging technologies, like single cell imaging and genomics platforms.



010. What's something that you don't enjoy about current bioinformatics research?

Computing in the cloud is promising, but it is still to expensive to store massive data for ongoing active compute and too slow to move data into the cloud and out again for each analysis.



011. If you could go back in time and visit yourself as a 18 year old, what single piece of advice would you give yourself to help your future bioinformatics career?

Keep taking math classes.



100. What's your all-time favorite piece of bioinformatics software, and why?

The UC Santa Cruz Genome Browser: you cannot underestimate the importance of looking at raw data, and the browser provides a way visualize a lot of data for every position of the genome. It is easy to check if your assumptions are right or not.



101. IUPAC describes a set of 18 single-character nucleotide codes that can represent a DNA base: which one best reflects your personality, and why?

S for strong.

An automated attempt to identify duplicated software names

From time to time I've been pointing out instances of duplicated software names in bioinformatics. I assume that many people reuse the name of an existing tool simply because they haven't first checked — or checked thoroughly — to see if someone else has already published a piece of software with the same name.

I am not alone in my concern over this issue and Neil Saunders (@neilfws on twitter) has gone one step further and written some code to try to track down instances of duplicated names. He recently wrote a post on his What You're Doing Is Rather Desperate blog to explain more:

In the article, he describes how he used a Ruby script to parse the titles of articles downloaded from PubMed Central and then feeds this info into an R script to identify examples of duplicated software names. The result is a long list of duplicated software names (available on GitHub).

I'm not surprised to learn that generic names like 'FAST' and 'PAIR' appear on this list. However, I was surprised to see that in the same year (2011), two independent publications both decided to name their software 'COMBREX':

BUSCO — the tool that will hopefully replace CEGMA — now has a plant-specific dataset

With the demise of CEGMA I have previously pointed people towards BUSCO. This tool replicates most of what CEGMA did but seems to be much faster and requires fewer dependencies. Most importantly, it is also based on a much more updated set of orthologous genes (OrthoDB) compared to the aging KOGs database that CEGMA used.

The full publication of BUSCO appeared today in the journal Bioinformatics. I still haven't tried using the tool, but one critique that I have seen by others is that there are no plant-specific datasets of conserved genes to use with BUSCO. This appears to be something that the developers are aware of, because the BUSCO website now indicates that a plant dataset is available (though you have to request it).

2015-09-21 at 9.42 AM.png

Anatomy of an mainstream science piece →

A great blog post by Ewan Birney that describes the process of writing an commentary piece for the Guardian newspaper, and which also discusses the need for more involvement of scientists in the public discussion of science. I like the concluding remarks:

As practicing scientists, we need to continue laying the groundwork started long ago by many others…engaging consistently and non-judgmentally with our communities and policymakers about out work. There is a real task ahead of us in providing an accessible way for people to digest this information. We should take every opportunity to communicate on every level, from the most basic to state of the art. Only then can society really use the hard-earned information gleaned from genetics appropriately, and for the greater good.

A long time ago in a Galaxy browser far, far away…

Today marks exactly 10 years since the first publication that describes the popular Galaxy platform:

The tremendous success of the Galaxy project can be summarized by the following two graphs (taken from the statistics page on the Galaxy Wiki):

New user registrations on Galaxy main site

 

Publications that reference or mention Galaxy

I'm sure that the Galaxy team have a more official date to use as their anniversary, but I'll mark the 10th year since their initial publication to say congratulations and 'Happy Birthday'! I hope that Galaxy can emerge unscathed from those difficult 'teenage years' that lie just around the corner!

Twitter competition: win a signed copy of Bioinformatics Data Skills by Vince Buffalo

I have an extra copy of the fantastic Bioinformatics Data Skills book by Vince Buffalo (who you should all be following on twitter at @vsbuffalo). I've come up with a fun little competition to let someone have a chance of winning this signed copy.

All you have to do is write a tweet that includes the #ACGT hashtag (so I can track all of the answers), which provides the following information:

Name a useful bioinformatics data skill

The winner will be chosen randomly — hopefully by a powerful scripted solution that Vince will help me with — in two weeks time. I will post details of any interesting (or funny) suggestions that you come up with on this site. The full details are below.

Competition rules

  1. Tweets should name a 'useful bioinformatics data skill'
  2. Tweets must contain the hashtag #ACGT
  3. Last day to enter into the competition is 25th September (midnight PST)
  4. One winner will be chosen randomly
  5. Only one entrant per twitter account
  6. Retweets of tweets that use #ACGT hashtag will be excluded

Trying to locate the source of duplicated software names

Thanks to Andrew Su (@andrewsu) and Mick Watson (@BioMickWatson) for alerting me to the following:

The former paper is from 2009, the latter paper is from 2015. Neither paper has anything to do with this 2010 paper which introduced something called the Genome Positioning System (GPS). Most importantly, none of these papers have anything at all to do with GPS (as most people understand the term).

If I run a Google search for GPS Bioinformatics the top hit that I see is for the MSc course in Bioinformatics as part of Brandeis University's Graduate Professional Studies program.

The usual disclaimer applies:

  1. Check existing literature before you name your software (at the very least run a Google search).
  2. Double check the name by adding the word 'bioinformatics' or 'genomics' to the search terms.
  3. Avoid names which wholly or partially contain words or terms that have nothing to do with your software.

Freely & Unrepentantly Confessing to Heresy →

There is a new post by Keith Robison in which he comments on my Excel/Bioinformatics post from August 28th:

Keith Bradnam reported a huge influx of traffic for a recent post -- not surprising, since he labeled it NSFW (Not Safe For WorK)… Bradnam was, of course, kidding. His short item showered derision on a recent Microsoft announcment about importing sequences into Excel.

The spike in traffic really has been insane. The post has become my most viewed post of anything I have ever written on this blog (by quite a margin). Clearly I have tapped into some anti-Microsoft (or just anti-Excel?) sentiment.

Keith Robison then takes on the 'case for the defense' and makes some fair points about Excel:

But to dismiss Excel as unworthy of any use in bioinformatics is to miss the fact that buried under the residue of years of creeping featurism is a tool useful in specific contexts and with some key advantages. The first advantage is that it is ubiquitous…

He then goes on to include some legitimate examples of how you might want to use Excel in order to work with sequence data.

I will conclude by saying that I bear no ill feelings towards Excel users, even those using it for bioinformatics! I have also used Excel in the past for trying to analyze some bioinformatics data. This was using Excel 2004 for Mac which suffered from a 32,000 row limit which made it unsuitable for incorporating some datasets. It was this limitation that was partly the impetus for me to start learning how to do some things in R.

Fundamentally, I feel that Excel is not a great tool for bioinformatics because: it is not open, it is not obviously workable as part of any standard bioinformatics pipeline, and it is not available on Linux (where a lot of bioinformatics happens). But, as always, you should use the tools that help you get the job done.