I’m pleased to turn the CSUPERB blog over to Bori Mazzag (Humboldt State University) and Sandra Sharp (CSU Los Angeles)! Bori and Sandy are both members of the CSUPERB Strategic Planning Council. They jointly organized the “Preparing Students in Bioinformatics: Challenges, successes and opportunities” workshop held January 8, 2016, during the 28th annual CSU Biotechnology Symposium at the Hyatt Regency Orange County. Here is their workshop report:
This year for the first time, the CSUPERB GATC (Genomic Analysis and Technology Committee) and the QB (Quantitative Biology) faculty networks hosted a joint workshop at the 2016 CSU Biotechnology Symposium, entitled “Preparing Students in Bioinformatics: Challenges, successes and opportunities.” The two groups are led by Sandy Sharp (CSU Los Angeles, CSULA) and Bori Mazzag (Humboldt State University), respectively.
The GATC has focused on research-based curriculum since its inception in 2010. QB has worked on exploring and disseminating best practices in quantitative (mathematical, statistical and computational) preparation of CSU biology students. The two interest groups have organized separate sessions at the CSUPERB Symposium since 2012, but with members of each group expressing a desire for a workshop on bioinformatics, we decided to undertake a combined offering. It turned out to be a powerful, well-received strategy!
We concentrated on two main themes: (1) Does the current bioinformatics training offered by CSU campuses meet workforce needs? (2) What are some differences between the skills needed by end users/biologists versus programmers/software developers?
To address these questions, we invited a panel of CSU faculty, industry employers and hiring managers: Jamil Momand (Chemistry and Biochemistry, CSULA) and Nancy Warter-Perez (Electric and Computer Engineering, CSULA), Paul Wolber (Agilent and San José State University, SJSU), Soheil Shams (BioDiscovery, Inc.), and Pleuni Pennings (Biology, San Francisco State University, SFSU).
The “Preparing Students in Bioinformatics” workshop featured a “bi-directional” panel of CSU faculty, industry professionals and hiring managers. The panelists were (left to right): Paul Wolber (Agilent and San José State University), Soheil Shams (BioDiscovery, Inc.), Pleuni Pennings (Biology, San Francisco State University), Nancy Warter-Perez (Electric and Computer Engineering, CSU Los Angeles) and Jamil Momand (Chemistry and Biochemistry, CSU Los Angeles).
The ~45 workshop attendees included faculty members from 18 of the 23 CSU campuses and 4 California Community Colleges. Faculty from a wide range of disciplines participated, from Biology, Chemistry and several different types of Engineering. Participating faculty members either aspire to or are already involved in bringing bioinformatics into the classroom, often as components of original student research. The workshop facilitated the sharing of expertise and challenges in genomics/informatics research and teaching, with the aim of providing the best possible learning experiences and outcomes for our students.
The workshop began a panel discussion, followed by round-table discussions among the participants, with panelists and other faculty participants serving as discussion leaders. Near the close of the session, each table shared a few important lessons learned. Workshop participants who had curricular materials to share also spoke briefly about their work and the materials they were offering.
First, each panelist described his or her training and current work in genomics/informatics research and education. It was fascinating, and even somewhat surprising, to realize that each panelist arrived at their current position by career paths that required them to cross disciplinary or sector boundaries. For example, Paul had long been working at Agilent when he became an external evaluator and advisor for the SJSU bioinformatics program. The collaboration grew, and Paul has now been teaching an engineering course on next-generation (“next-gen”) sequencing technologies at SJSU for several years. He also described positive impacts his teaching job had on his own industry career.
Drs. Warter-Perez (left) and Momand (center) describe a biology-computer science course they co-taught. Bori Mazzag (right, Humboldt State) moderated the panel discussion.
Jamil and Nancy gave a detailed description of the bioinformatics course they team-teach to computer science and biology students. They shared strategies and philosophies around designing a successful course that meets students at their current level, but also pushes them to complete an interdisciplinary project. As instructors from very different backgrounds, they model the sort of interdisciplinary collaboration they ask from their students in the course.
Soheil talked at length about the need for programmers at his company and the pressures his and other, similar mid- and small-sized bioinformatics companies feel in competing with the likes of Google or Uber for software engineers. From his perspective, one of the greatest challenges is to inspire (or recruit) and train (and retain) programmers who are willing to forego some potential earnings to work on and solve the new and impactful problems that biology has to offer.
Pleuni shared information about her own training that built on a biology background but became increasing quantitative. The bioinformatics course she developed draws upon that experience.
The academic panelists spoke on skills or knowledge they want to impart to students and what challenges they face as instructors with students varied levels of preparation. Industry representatives talked about important skills needed by programmers and end users, but seen as lacking in current job candidates. Not surprisingly, both written and oral communication were mentioned as a critical skill for all students entering the job market. Several examples were given of the specific types of writing required and it was clear from Paul and Soheil’s comments that although writing is more emphasized in teaching STEM fields nowadays, there is a need for continued improvement in this area.
A discussion developed around the term “bioinformatics” itself. Soheil argued that it is too broad an umbrella term that encompasses vastly different types of jobs and skills.* This point resonated with many in the audience and came up as an important takeaway in the post-workshop survey. As it often happens in interdisciplinary settings, the panelists and audience were split on the relative importance of increasing depth of knowledge in a discipline versus interdisciplinary training. Often those who advocate for depth see interdisciplinary projects primarily as a way to teach students how to communicate and work with others from a different discipline. Others see interdisciplinary training as a fundamental, new approach to teaching content to students. This perspective often implies that the current academic silos do not provide the right organizational structure for the sort of work that graduates do or problems they face after leaving school.
As CSU faculty members, it was interesting to us to discover the diverse ways campuses, departments and researchers have devised to teach bioinformatics and mentor their students in this field. Because bioinformatics is so new and technologies evolve so quickly, professionals and professors in the field must stay agile. For some, it means learning new programming languages, building new collaborations around research questions, pooling resources, or sharing equipment across campuses.
One challenge for faculty has been finding resources and a network of CSU researchers in the field, dispersed in home departments across engineering, computer science or biology. CSUPERB, and the GATC specifically, has been instrumental in facilitating networking across the CSU. Bay Area and Southern California campuses certainly benefit from having tight connections with industry (for example, the SJSU/Agilent relationship). But far-flung or isolated campuses, such as CSU Chico, have successful bioinformatics concentrations as a result of purposeful hiring and faculty networking.
The panel discussion was followed by round-table discussions involving all participants. One table gathered faculty who all teach a bioinformatics course. The course designs described varied greatly. Much of the discussion centered on recruiting students to take bioinformatics courses (for whom programming is often a deterrent) and how to teach programming most effectively. Most agreed that for biology courses in which informatics software is used as a tool, it is necessary to treat the software itself as a “black box” to some extent, but there was disagreement over where the right limit was.
An interesting side-note: MANY participants wrote they would like to see further workshops and training on programming, either in R or Python or other languages. CSUPERB will look into inviting Software Carpentry to offer a workshop to faculty this year.
Janey Youngblum (CSU Stanislaus) motivates students by getting genomes sequenced by 23 and Me. This resonated with many of the participants. While there are several problematic, ethical and financial considerations around using students’ genomic data, making research questions personally engaging and relevant clearly struck a chord among the workshop participants.
Amy Sprowles (Humboldt State) takes notes as Judy Brusslan (CSU Long Beach) describes CyVerse as a bioinformatics teaching and learning platform.
Two other tables had extensive discussions about programming and software. Clearly, finding the right software to teach at the right level has been a huge obstacle to effective teaching and desired learning outcomes. As more on-line tools become available, and some methods standardize, this hurdle may become easier to cross over time. One of the most animated table discussions was on iPlant (or now CyVerse). The discussion morphed into a mini-tutorial by Judy Brusslan (CSU Long Beach) on how to use CyVerse for the analysis of large data sets. Because CyVerse has cloud-based data storage and analysis tools, many concerns about availability of computational power and storage space have been alleviated.
Several members of the audience, including David Keller (CSU Chico), Jose de la Torre (SFSU) and Renaud Berlemont (CSU Long Beach), talked about their curricular materials and shared materials or posted links to them. We also learned that Jamil and Nancy’s book, based on their team-taught course, is going to be published this year.
In the immediate post-workshop survey, a handful of main takeaways were mentioned. These open text replies are representative:
- “There are (effective) strategies to teach bioinformatics to bio students without programming skills”.
- “Next-Gen Analysis is possible at CSU.”
- “Finding out what’s going on related to bioinformatics on other campuses.”
- “Learning about different resources to use in my courses.”
- “Lots of cool resources and programs at different institutions.”
The session also received very positive feedback from faculty participants who responded to the post-symposium survey. We have been in contact with the workshop attendees, disseminating workshop notes, contact information and curricular materials. There is a lot of energy around bioinformatics (for lack of a better term!) networking and developing a workshop around hands-on skills (perhaps programming!) next year.
— Bori Mazzag & Sandy Sharp
*Editor’s note: A recent search of bioinformatics-related corporate (non-academic) job position openings in California (Jan. 2016) used these technical keywords: statistics, data analysis, data management, quality control, version/revision control, software engineering, methods development, algorithm development, expertise in a series of scripting languages (“Perl or Python, R, shell scripting, and MySQL”), programming skills in a series of languages (“Python or Perl/Java/C++; shell scripting and R”). Many job descriptions dedicate more text to “soft skills” (communication, agility, teamwork, etc.) than technical skills.