By Genomics Aotearoa 29/07/2020

By Associate Professor Mik Black, Department of Biochemistry, University of Otago

The increased availability of complex biological data sets means that analysis and computation are becoming critically important skills for New Zealand’s future scientists. Because of this, we need to be doing everything we can to help our students develop these skills, to better prepare them for large-scale data analysis across a range of fields.

At present, students focusing on the biological sciences in New Zealand tend not to take the more “computational” subjects – statistics, mathematics, computer and/or information science – yet the growing importance of data manipulation and analysis in genomics make these important and sought-after skills.

While there are now more students coming through biology-focused programmes like genetics and biochemistry who are starting to include computational/analytic training (great to see), those numbers need to grow.

What is driving the need for data analysis capability in genomic science?

Mik Black

Over the past two decades, the field of genomics has moved at an incredible pace, and the volume of biological data we are able to produce with high throughput methods is staggering.

There is now widespread acceptance that genomic technologies can help us to understand the world around us, and in particular improve human health.

It also has huge implications for conservation – genomics has the potential to play an important role in helping to slow the global decline in biodiversity.

While it is relatively easy to generate large amounts of genomic data about any organisms, it’s what you do with the information that is important.  And this is what has fuelled the strong demand for analysis and interpretation, putting disciplines like Bioinformatics and Data Science in hot demand.

Bioinformatics is the development of methods and software tools for understanding the biological data derived from genomics. It is an evolving science, but is essential to manage data in modern medicine and biology.

For me, as a statistician, Data Science is like statistics on steroids. While everyone has a slightly different definition, it is largely a blend of key skills from mathematics, statistics and computer/information science, focusing on the management, manipulation, visualisation and analysis of large and sometimes disparate data sets.

The combination of these skill sets – bioinformatics and data science – is what will help us address the considerable computational challenges involved in developing pipelines that take raw biological data as input, and output practical tools for use by health professionals and those at the front lines of primary production and conservation.

While New Zealand definitely has high-quality people that are skilled in these areas, we need a whole lot more to help us find solutions to some of our country’s major challenges, such as health inequities.

A comprehensive approach to building this capability will allow us to accomplish many things, from implementing precision medicine models to help doctors choose treatments that are a genetic match for their patients, to developing practical applications to help save endangered species like the kākāpō.

So, what are we doing?

 The government has recognised the importance of genomics to New Zealand’s future. It has funded Genomics Aotearoa to build bioinformatics capability, to make use of genomics research already happening internationally, and to use New Zealand-specific genomic information for good in our own country.

A planned programme of training within Genomics Aotearoa is aimed at retrofitting these skills for New Zealand genomics researchers. We have already started programmes to train people who can then go own to teach tools to people in their own organisations.

In partnership with the New Zealand eScience Infrastructure (NeSI), we have been delivering Carpentries training ( to teach generic skills in programming, version control and reproducibility to researchers who have not been exposed to these concepts as part of their formal education.

We are now using the Carpentries model to deliver training for bioinformatics skills. Genomics Aotearoa was one of the first in the world to run the Genomics Data Carpentry workshop, delivered to almost 200 New Zealand researchers in its first six months, and more workshops are on the way – Metagenomics, RNA-sequencing, Genotyping-by-Sequencing, Reproducible Research for Genomics – all free and open source.

We are already making a huge difference for students and researchers. But we also have a responsibility as scientists to ensure that school and university students are aware of the importance of data analyses in all science subjects, and are being encouraged to develop these skills.

Bioinformatics and Data Science skills need to be seen as a fundamental component of modern genomic science. Universities can still do more to incorporate these skills into our existing papers, and in the longer term, we would like to see courses teaching these skills being integrated into university degree programmes to reflect their growing importance across the entire science research spectrum.

0 Responses to “Data analysis skills are in hot demand – what should we be doing about it?”

  • There are institutional things hold back a strong workforce in this space. Most data scientists currently are either former academics working in professional roles or professional computer scientists (or similar) working in semi-academic roles. These people are expected, and expect, to do research work, supervise research students and so forth. But the roles are often cost-recovered professional roles that give no weight to publishing papers, supervising students or other academic expectations. Further, there is no room for these people to grow in their roles. They need to be paid well, because they are expected to be highly trained (often with a huge student loan debt) and highly sought after. So to date, they’ve gone to work in institutes overseas, or more recently to private industry. Academic institutions need to do more than train people – they need to give trained people stable jobs that pay well and have a future, so that they progress themselves, rather than being trained and then forced to leave.

    • Mik Black replies …This is certainly true in an academic setting, where we tend to be tied to traditional ideas of discipline-defined research groups led by a senior expert from a specific field. In this sort of framework, individuals involved in data analysis (no matter how skilled, experienced or talented) are often relegated to “support” roles, without long-term job security. There are moves internationally to change this situation, however, and in particular I would like to point to the excellent work being done in this space by NeSI, who are strong advocates for the Research Software Engineering (RSE) community, which also faces these challenges (and is broad enough to encompass Data Science practitioners). Shameless plug for the upcoming (free) online New Zealand Research Software Engineering Conference:

  • Hi Mik – my experience too (rather painfully). It’s doubly frustrating if you come from a very research-oriented focus, and have projects you’d like to tackle. In can be compounded by that—just an opinion!—NZ isn’t a good country for projects that naturally try look a fair way ahead, as funding is focused on short-term results.