Career Outlook: Data and Bioinformatics Scientists (2018-26)
The Future of Jobs
The fourth industrial revolution is creating many new opportunities as companies increasingly leverage new technologies by rapidly transforming in a major way the nature of human labor by augmenting through technology and assigning jobs to machines and computer algorithms.
- The Future of Jobs
- The top ten job winners
- The top ten job losers
- What is the career outlook for Bioinformatics (2018-2026)?
- What skills a successful bioinformatician need to have?
- Is bioinformatician a jack of all trades, master of none?
- Specialization within a branch of bioinformatics is essential
- How many vacancies are there currently?
- Not all bioinformatics jobs and bioinformatics scientists are the same
- Job growth and median wages
- Keep an eye on rapid technology development
- What technology knowledge employers demand today?
- Formal bioinformatics education
- Pros and cons of bioinformatics as an interdisciplinary field
The exponentially rapid technology development will increasingly discharge manual work needed mainly from data processing, information acquisition, and likely more commonly support high-value responsibilities such as data analysis, reasoning, and various decision-making tasks.
The Future of Jobs Survey 2018, World Economic Forum, estimates that during the period of 2018-2020 mobile internet, big data analytics, artificial intelligence, and cloud technology are the dominating trends and drivers for business growth. For example, 85% of surveyed companies are likely or very likely to adopt big data analytics.
World Economic Forum survey estimates that the industrial revolution will create 133 million new job roles and that 75 million jobs will be disappearing by 2020.
According to the estimate in the report predicts that even a moderate adoption of new technology results in up to US$8 trillion investment over the next two decades in the US alone.
The top ten job winners are:
- 1. Data Analysts and Scientists
- 2. Artificial Intelligence and Machine Learning Specialists
- 3. General and Operations Managers
- 4. Big Data Specialists
- 5. Digital Transformation Specialists
- 6. Sales and Marketing Professionals
- 7. New Technology Specialists
- 8. Organizational Development Specialists
- 9. Software and Applications Developers and Analysts
- 10. Information Technology Services
The top ten job losers are:
- 1. Data Entry Clerks
- 2. Accounting, Bookkeeping, and Payroll Clerks
- 3. Administrative and Executive Secretaries
- 4. Assembly and Factory Workers
- 5. Client Information and Customer Service Workers
- 6. Business Services and Administration Managers
- 7. Accountants and Auditors
- 8. Material-Recording and Stock-Keeping Clerks
- 9. General and Operations Managers
- 10. Information Technology Services
The report lists increasing protectionism, cyber threats, shifts in government policies, the effects of climate change, and aging societies as negatively affecting business growth.
What is the career outlook for Bioinformatics (2018-2026)?
The simple answer to this question is that the overall outlook is excellent, the demand outweighs the supply, but the devil is in the details as usual. Nevertheless, it is good to be a bioinformatics scientist.
First and foremost it is the question about specific skills. Bioinformatics is an interdisciplinary field; Thus, relating and interacting with many other areas of knowledge. People practicing bioinformatics originally come from a wide variety of disciplines, and therefore their specific skills also vary widely.
There are many jobs available, but the required skills also vary enormously. It is not difficult to find bioinformaticians, but it is challenging to find the ones that possess the specific skills for a particular job. Why is it so?
Let’s take the most confusing part away first. When we look at the definition of bioinformatics “the branch of science concerned with information and information flow in biological systems, esp. the use of computational methods in genetics and genomics,” and how it turns out to look like when we browse job advertisements (See definition of bioinformatics).
What skills a successful bioinformatician need to have?
A small survey of job advertisements that come up with the keyword ‘bioinformatics,’ reveals that bioinformatics is related to a wide range of qualifications:
The ads specify a required degree within Bioinformatics, Bioengineering, Computational Science, Software Engineering, Machine Learning, Mathematics, Statistics, Molecular Biology, Biochemistry, Computer Science, Biostatistics, Biomedical Engineering, Engineering, Biology, Information Systems, Genomics, Computational Biology, or related field.
The specific degree level varies. Within industries, the requirement ranges from Bachelor’s, Master ’s to Ph.D., whereas in academics a Ph.D. is usually required.
No single advertisement lists all of the fields but often do include "or related field" at the end of the list; Thus, the list emphasizes the interdisciplinary nature of bioinformatics and the fact that surely no single person exists who can master the whole range.
Consider that biology consists of more than 30 over sub-disciplines and branches of which many of them we can further divide into more specifics, e.g., molecular genetics, classical genetics, genomics, functional genomics, and population genetics, etc. (Example list of sub-disciplines in biology).
Is bioinformatician a jack of all trades, master of none?- Alternatively, is a bioinformatician master of one, knows a little bit of another one?
The brief answer is 'none of the above.' Bioinformatics is an interdisciplinary field, and thus the requirement is to be a master of at least two separate areas.
Back in the early days when formal education programs in bioinformatics were not available, people taught themselves often driven by a passion and also for the clamor given to the 'magicians' by biologists who didn't know anything about programming.
On the other hand, one could also receive less encouragement like me back in the 1990s when I was a Ph.D. student at a medical genetics department, one of the professors told me that "you cannot do biology in computers." OK, I and thousands of others have proven him wrong by now.
The future bioinformaticians were people either with a biology background who taught themselves programming or with a computing-related background who taught themselves biology.
Yes, there was a barrier of separate disciplines using different vocabularies, but passionately driven people were able to break the barriers perhaps because they were working with people with specific narrow interests; Thus, the amount of biology needed to learn was comfortably limited.
On the other hand for a biologist, the threshold to learn programming was also lower due to the relative simplicity of the tasks compared to what expectations are today.
The rapidly increasing amount of data drove the transformation from biologists or computer scientist to bioinformatics scientists. In the early days, the essential everyday tasks were mostly to do database searches and related sequence analyses.
Times have changed and keep rapidly evolving. Bioinformaticians are not considered to be magicians anymore. Definitively simple database searches and some sequence analysis and even many more advanced analyses using current software tools are functions that are routine today, and a non-bioinformatician biologist is expected to be able to handle alone without help.
The main point is that every successful bioinformatician has his or her particular specialization. There is no more trustworthy know-it-all magicians or gurus. Specialty is not bioinformatics but a branch within bioinformatics!
Specialization within a branch of bioinformatics is essential
How many vacancies are there currently?
The first question was, how many jobs are there? To try to answer that, I used LinkedIn's job search with the keyword 'bioinformatics' that found a total of 7,853 vacancies worldwide. Glassdoor found 10,226, Monster 1,954, Ziprecruiter 1,918, and SimpyHired 2,504 vacancies.
Most of the jobs seem to be in the US and EU, except for LinkedIn which found 2,444 vacancies in India. A 3,459, 8,761, 1,954, 1,918, and 2,504 were in the US found by LinkedIn Glassdoor, Monster, ZipRecruiter, and SimplyHired respectively.
I did these searches on October 7th, 2018, but have done similar searches several times earlier and seen that most of the vacancies are in North America and Europe. I believe there is much truth in the bias, but perhaps some of it is due to the location and language of the search platforms. Besides, this only a tiny sample of everything there is available for job search.
Many employers advertise many vacancies only locally and in a local language. Besides, a high portion of jobs never ends up advertised along with the fact that many get filled through acquaintances. One note here though, some advertisements are for jobs that are not vacant anymore to fulfill legal requirements. See an article in Nature on September 19th, 2018 "Job vacancies posted after being filled: it’s time to stop wasting everyone’s time."
Not all bioinformatics jobs and bioinformatics scientists are the same
As I stated above, it is hard to find a suitable bioinformatician even though any vacancy advertisement often attracts thousands of applications, emphasizing the fact that the demand is still higher than the supply.
Because professional bioinformaticians have a wide variety of different specializations, finding a particular specialty that is ideal for a specific job is still often fruitless endeavor. Open positions still often get filled with candidates who are less than perfect, and applicants are thus expected to learn along the way. I guess that the phenomenon applies to much other employment types, but that is beside the point here.
Please misunderstand me correctly this does not apply to the entry-level employment, where employees are meant to learn the trade by doing the work.
The difference is that for example a laboratory technician position only requires a bachelor's degree whereas an academic post necessitates the completion of a doctoral degree. However, often a bachelor's or master's degree is sufficient to work within the industry as we see in the vacancy advertisements.
Job growth and median wages
O*NET OnLine, sponsored by the U.S. Department of Labor projects the job growth for bioinformatics scientists nationwide to be 5 to 9%, and in California as high as 12%. They project the total job openings to be 3,700 during the period 2016-2026 with the total employment in 2016 being 39,000 employees.
For bioinformatics technicians, the projected growth is the same 5 to 9% percent as for scientists and the total employment 12,000 employees in 2016.
According to O*NET OnLine, the median wages in 2017 were US$76,690 annually for bioinformatics scientists and US$47,700 for technicians.
On European and Asian levels similar statistics are cumbersome to find because they have to be collated country by country. However, presumably, the growth will follow the overall global trends.
In any case, median salaries give only a vague idea of reality since the wages vary enormously between levels of employment.
The number of vacancies in the future is only one side of the coin. We need to look at the number of bioinformaticians that graduate during the same period. According to GEEBEE, 27 universities offer Ph.D. programs in bioinformatics.
If each of the programs produces an average of about 14 graduates each year, it results in about 14*10*27 = 3,780 bioinformatics scientists, which is about the same number as the predicted vacancies during the ten year period. How many will graduate in reality? Your guess is as good as mine.
Then there is the fact that many bioinformatics scientists will retire during that period. How many? I don't know that figure either, but the retirement will increase available job opportunities. So, at least at the moment of writing there will not be any massive overproduction of fresh graduates.
Keep an eye on rapid technology development
Both laboratory and computing technologies are constantly and rapidly evolving. Let's look at one of the prominent examples, sequence assembly. Back in the 1990s, the human genome project used yeast artificial chromosomes (YACs) to first clone portions of the genome sequence and subsequently sequencing each YAC. After sequencing each YAC had to be mapped to the genome. However, YACs have a high tendency to recombine.
Then came bacterial artificial chromosomes (BACs), but the requirement to map them onto the genome remained.
A game changer in genome sequencing came with the shotgun method in which the entire genome is cut into short random pieces and then assembled; Thus, eliminating the time-consuming mapping step.
In the early days, the Sanger sequencing method was the method of choice and produced sequences that were from 400 to 800 bases long. However, Sanger sequencing was expensive and still is today.
In an attempt to reduce sequencing cost, the next-generation sequencing (NGS) methods started to emerge that were able to produce a massive amount of sequences. This far so good, but new problems arose, for example, in the beginning, the read fragments were very short and the error composition very different from Sanger sequencing method.
Steadily, the read length increased and also new technologies emerged that were able to produce read lengths consisting of tens of thousands of bases.
All the laboratory technology development required rapid development of novel algorithms.
The completion of the human genome and concurrent development of both laboratory and technology methods resulted in data tsunami, yielding new opportunities requiring novel technologies.
Over ten years ago, in the mid-2000s Roger Mougalas coined the term Big data, a name that is now in everyday use. However, it turned out that relational databases are not efficient in handling big data; Thus, new technologies were again required.
A large amount of data are also changing the way how we apply machine learning techniques. In the late 1950s, when Frank Rosenblatt described the 'perceptron,' not much data was available, and therefore artificial neural networks were not as efficient as the 'traditional' machine learning algorithms, but all started to change after the turn of the millennium when large amounts of data became readily available from varied disciplines together with increased computing power.
"When computing technologies mature and cease to be novelties and become routine with accompanying user-friendly software, the tendency is that non-bioinformaticians are expected to use them without expert help."
Today, artificial neural net algorithms have been proven to be more accurate and versatile than other machine learning methods with deep learning, convolutional networks, reinforcement learning and so on.
The point is that you need to keep continually learning and perhaps developing novel technologies; otherwise, you'll be left behind and replaced with bioinformatician having current knowledge.
When computing technologies mature and cease to be novelties and become routine with accompanying user-friendly software, the tendency is that non-bioinformaticians are expected to use them without expert help.
Carefully observe where the data is that is most likely to result in novel discoveries and subsequent novel technologies. Data are the source of everything, and without it, nothing much will happen.
What technology knowledge employers demand today?
There is a huge demand for scientists who can analyze mountains of diverse data, make sense of it and be able to present findings in a clear and meaningful manner to decision makers.
Again, let's briefly look at what technologies the job advertisements contain. Here is a short example:
NGS, ChIP-seq, GWAS, eQTL, ATAC-seq, RNA-seq.
Cloud computing: HPC, Slurm, AWS, Azure, GCP, SGE, Docker, web service, and of course big data, machine learning, and statistical methods.
Grid technologies: Sun Grid, Torque, PBS.
The above is only an incomplete snapshot of employers' specific wish lists but overall seems to imply the tendency towards analysis of large amounts of data and big data, as already mentioned in the beginning in this article.
Then there is medical informatics and precision medicine or personalised medicine as it is called in Europe, but that must be another article.
In summary, bioinformaticians armed with cross-disciplinary skills and knowledge are surely in the forefront to tackle future challenges.
Formal bioinformatics education
Skill sets required for bioinformaticians are constantly evolving and growing. Long gone are the days when a self-taught bioinformatics expert could get a decent job. Today, a formal degree is a prerequisite.
A formal degree in bioinformatics provides a diverse set of skills that opens a range of career options.
Formal education structure can look as in Infographics 1, which is the formal education structure in Uppsala University in 2018, one of the top universities in the world.
Big data in healthcare spearheaded by many public initiatives is growing at an accelerated pace, giving opportunities within medical informatics, such as precision or personalized (UK: personalised) medicine. The emerging field of research aims to revolutionize the way we prescribe and develop medication. However, this is a topic for another article perhaps soon. Keep an eye on the blogs section.
Pros and cons of bioinformatics as an interdisciplinary field
Pros: Having a formal education, and thus a wide range of skills in biology, computer science, and the ability to handle big data, opens many opportunities even outside the life sciences around the world. You might be a part of the global talent pool.
Bioinformatics is challenging because you must master an interdisciplinary field. Biology, which itself is complicated and computer science and both having separate large vocabularies.
Some of my former Ph.D. students chose to embark on an industry career instead of joining the academia. One of them entered a large financial institution and a few others companies developing software.
Cons: The employment landscape and requirements are rapidly changing, requiring life-long learning to keep up with the latest technologies in both laboratory and technology sides.