Hale Ann Tufan is an Adjunct Assistant Professor of Plant Breeding and Genetics at Cornell University. She is also a member of the CGIAR Gender and Agriculture Research Network’s Gender and Breeding working group. Here she tells us a little bit about how she started to work in this area, what she’s working on right now and some challenges and potential opportunities of integrating gender into breeding.
A molecular biologist by training, Hale’s path to working on gender was not a direct one…
“After having worked in a lab, I realized I wanted to work around people. The transition from being a molecular biologist to other fields however, was not an easy one. I started looking at crops to work on that had a more applied focus, so cassava came out as a crop to work on in sub Saharan Africa. I got my PhD in the UK, and during a postdoc working on cassava I came across a job ad for NextGEN Cassava project manager. That’s how I got into more applied and development focused work, moving into this management position. The people who had designed the project had built in a gender budget into the project but there were no specific details on what had to be done. It was really exciting because it was a whole new world, and obviously gender issues in agriculture was interesting to me, outside of my work, but I had never had put those two together, and all of a sudden, it was that I could do this on a day to day basis.”
The winding road from molecular biology to gender and breeding
Q: Tell us how you came to work on gender work
A: What gender work within the NextGEN cassava project looks like is still being worked out but the initial idea was that it had to be applied to breeding, given the project’s objective. So, as a team there’s been a lot of thinking through and working towards how to apply gender integration into breeding programs. There’s a lot that has already been done. The sex disaggregated trait collection seemed to be the first entry point. This was basically how I got into the gender work, and working with really incredible people along the way just got me more and more interested.
Currently, I am part of a new project from the Gates foundation called GREAT, where we’re conducting gender training for biophysical scientists. There’s such a niche in working with scientists themselves, because there’s a lot of understanding and sensitization. However, it’s also sometimes quite unclear to scientists how to put this training to use.
There are some breeders for instance who are interested in integrating gender into their work but don’t know what to do with that interest, so that moved me more towards the capacity building end of things. I will still be spending about half my time doing research with NextGEN, however, I will also be a part of the capacity building program to design an optimum curriculum for biophysical researchers.
How to deliver these concepts so that they have that transformative “aha” moment, so that they are forever, on this side of the fence- that can be challenging. However, once they’re convinced, there’s no going back. They are very committed in including women in the research process or thinking about the impacts on women. So, that’s where I’m at right now.
Q: De-mystifying gender for breeders: what are some of the basics?
A: It’s an advantage to come from a research background because I completely understand their viewpoint, when the breeders have concerns or voice opposition.
Much of this disconnect is owing to the different viewpoints of being a breeder versus a social scientist. The world is a different place depending on your background, take for instance, your concept of valid data. How do you convince a breeder that the data is valid? What kind of data should it be and how should it be packaged? Do you need 2000 data points or very clear, black and white information? Is that what’s going to convince breeders or can you come up with an anecdote or story, more qualitative work?
It has been very interesting to see how we can take qualitative work- the work that we did in Nigeria is very interesting, but not at all qualitative. So how do we package that and how do we make that into something that convinces breeders, convinces them that it’s valid information. This is where the end user preferences piece comes in; that translation piece. Can we somehow have proxy traits, so if a respondent is saying that a particular crop cooks very well or that it lasts a long time in the soil, what does that mean? The types of questions we then ask are, “How long?, What kind of soil?.”
These are really the types of conditions that breeders can actually do something about. If instead, if we just tell them that it cooks well, they can’t set up 2000 cook stoves on a breeding trial and cook all the different varieties, to see which one cooks well. That’s the deal breaker, when it becomes too expensive or time consuming. So, one solution is to look at correlations between what respondents are describing and lab based variables.
Having said that, it’s easier said than done, as it requires a lot of groundwork. This is what we’re trying to set up now- field trials of material collected from farmers’ fields that are coupled with descriptions of those materials and why they are/aren’t good for certain products. We then took that planting material back to the breeding station, planted it on station, and are going to analyze these varieties using the food science variables, so that if a food scientist looked at the same exact variety, analyzing it with the more scientific variables which they use to explain quality and say what is the profile of a particular variety within a specific community, we can then pin point those proxy traits that the food science lab on station can start measuring.
If we find the proxy trait, and measure it in a new variety, we can start to predict which varieties we think may suit, based on the profiles we’ve attained. We then confirm that by taking it back to the farmers. That’s the full circle, the big picture. We have material back at the breeding station from Uganda and Nigeria. Now, we have the complex task of correlating respondent responses with food science variables; that type of correlation is what we’re going to do probably in the next year.
Q: What are some of the bottlenecks?
High costs for food science analysis is a constant challenge. Sometimes, you can spend upto 400 USD per sample and no breeding program can afford to do this along large populations. On the other hand, this is the classic wet chemistry done in food labs. Right now, in parallel, we are developing hand-held near infra-red spectroscopy machines. What those do is allow you to scan a root and give you a rough estimate of those variables. While not perfect, this allows for much cheaper phenotyping. Whilst there needs to be a benchmark to the lab analysis, this helps us find some shortcuts (which are less expensive) to present these traits. Right now, the cost of phenotyping is a major bottle neck but it may not be for long.
The challenge with data collection is cost, data size and still capturing “reality”. Sometimes the information is so distilled, that it’s oversimplified. What we are attempting to do, is avoid looking at just the numbers. This can be quite tricky because we want to make sure that we are explaining the numbers, but we need to do so with the right kind of qualitative work and that’s the work that’s hardest to communicate. How do we communicate qualitative work to scientists? How do we bridge that interdisciplinary gap? This is nothing new but for breeding it’s not a domain that’s as developed as it could be. Take nutritionists for instance, they are much more interdisciplinary.
Note: this interview will be published in two parts. This is Part I. Part II will be published shortly.