New publication: The dynamic balance of import and export of zinc in Escherichia coli suggests a heterogeneous population response to stress

I am absolutely delighted to see the (online) publication today of our paper:

Takahashi H, Oshima T, Hobman JL, Doherty N, Clayton SR, Iqbal M, Hill PJ, Tobe T, Ogasawara N, Kanaya S and Stekel DJ 2015. The dynamic balance of import and export of zinc in Escherichia coli suggests a heterogeneous population response to stress. Journal of the Royal Society Interface DOI: .

The abstract is at the bottom of the post. First I want to say why I am so happy about this particular paper.

1. This is the first piece of work I have published in which I have made a successful funding application (I say “I” loosely, as Taku Oshima wrote the Japanese part of the bid along with Naotake Ogasawara, Shigehiko Kanaya and Toru Tobe, and Jon Hobman wrote much of the UK part of the bid; however, I was PI as this was a Systems Biology call); led the research (most of the hard work was done by Hiroki Takahashi and Taku Oshima); wrote the paper (different parts were written by different people – almost all authors made an important contribution); and have seen the paper published. A complete cycle of more than 5 years from grant application to publication.

2. This is the first paper on which I am corresponding author that contains new experimental results. More than that, I played an important role in devising many of the experiments (time courses and viable cell count assays) – not in terms of technical details (in which I have little expertise) but in terms of what experiments we need and why we need to do them.

3. This is the first time I have come a complete “Systems Biology” cycle: from experiments, to models, to predictions, to experimental confirmation, to a new model and then new predictions.

4. Some of the most important ideas in this paper came as a result of one of the most enjoyable weeks in my scientific career. Hiroki, Jon, Selina and I were all attending the Biometals Conference in Brussels in July 2012. While Hiroki and I attended a few sessions of the conference, most of the time we spent in our hotel lobby trying to get the model to fit the data. The first data we had were the LB data, which the model could fit without too much trouble. When Taku asked me what other data we needed, I said: “time series, with different concentrations of zinc”, and by the conference we had those data too. Only one problem: the model no longer fitted the data.

Hiroki and I spent that week trying to work out how to get the model to work. Each morning, we sat in the hotel lobby, and over endless cups of tea, we devised new versions of the model; each afternoon Hiroki would code them up, and then run them overnight on the supercomputer in Japan. And the next morning, the model would still not fit the data. By the last day we were tearing our hair out: nothing we could think of would work. We had one last (desperate) idea: ditch the 100uM zinc data. Boom! The model fitted the 12.5uM zinc data just fine! And this led to the heterogeneity hypothesis, the viable cell count assay, the stochastic model, and all the results that make this paper exciting. It is that kind of week that we go into careers in science for: the frustration and delight of grappling with and overcoming a difficult problem.

5. As alluded to in the previous points, it has been an absolute delight working with my coauthors on this paper, and I have many happy memories in the UK, Japan and Brussels.

6. Finally, on a more personal note: we received this funding on 17th March 2010. In between receiving the funding and having the paper published (5 years later) I have got married (July 2010), had a baby (July 2011), had another baby (November 2013) and am now more tired yet more happy than at any other point in my life 🙂

Now for the abstract!


Zinc is essential for life, but toxic in excess. Thus all cells must control their internal zinc concentration. We used a systems approach, alternating rounds of experiments and models, to further elucidate the zinc control systems in Escherichia coli. We measured the response to zinc of the main specific zinc import and export systems in the wild-type, and a series of deletion mutant strains. We interpreted these data with a detailed mathematical model and Bayesian model fitting routines. There are three key findings: first, that alternate, non-inducible importers and exporters are important. Second, that an internal zinc reservoir is essential for maintaining the internal zinc concentration. Third, our data fitting led us to propose that the cells mount a heterogeneous response to zinc: some respond effectively, while others die or stop growing. In a further round of experiments, we demonstrated lower viable cell counts in the mutant strain tested exposed to excess zinc, consistent with this hypothesis. A stochastic model simulation demonstrated considerable fluctuations in the cellular levels of the ZntA exporter protein, reinforcing this proposal. We hypothesize that maintaining population heterogeneity could be a bet-hedging response allowing a population of cells to survive in varied and fluctuating environments.

Gender Equality in Appointments and Promotions: Two Proposals

The aim of this post is to put forward proposals to improve equality in academic appointments and promotions. The primary intention is to improve prospects for women academics, although these proposals would also benefit both women and men who have taken career breaks, work part time or who have worked in industry or other sectors. I propose two ideas that are detailed below: first, a move away from ‘career total’ publication metrics in favour of ‘personal best’ metrics; second, a commitment to gender balanced panels.

I am putting these proposals for discussion within the school/university in which I work. However, the problems (and solutions) are really quite general – certainly in the UK and probably internationally. Thus I am putting these ideas into the public domain because I would like to see them discussed and implemented across all institutions. I have also taken a ‘life science’ focus because that is the domain in which I work; I am sure that there are similar challenges in other fields.


The graph shows a ‘typical approximate proportion’ of women and men at undergraduate, postgraduate, postdoctoral, assistant professor (lecturer), associate professor (senior lecturer / reader) and professorial level in a ‘typical’ Life Sciences department: these data are based on some real data but have been modified to give a straighter line and to preserve anonymity of the data sources.


While the majority of students are female, the majority of staff are male, and the proportion of women decreases at every stage of career development. There are many complex reasons for this; nonetheless, this situation is a disgrace, not to mention a tremendous waste of talent. It is incumbent upon us to take action to remedy this. There are likely to be very many actions that could help. Here are two I propose: they are certainly not exhaustive.

Personal Best Metrics

My first proposed change is to stop using ‘career total’ publication metrics for appointments, promotions and academic biographies and move instead to ‘personal best’ metrics. Career total publication metrics include total number of papers authored, total citations, h-index and other such measures. For example, a university might have guidelines for the number of publications needed for a professorial appointment. These metrics discriminate against people who have had career breaks (especially women), worked in other (relevant) sectors (industry or public bodies) or who work part time (e.g. 80%; also often women).

The ‘personal best’ metrics I propose have solid precedent and foundation, namely in their use for REF. Here is an example of how such a system might work for appointment or promotion:

  1. The judgement is made on a suitable period prior to the appointment/promotion. For example, 6 years, as per the REF, seems very sensible.
  2. The applicants’ top 4 research outputs in this period are considered. These need to be outputs where the applicant is a leading author (i.e. first, last or corresponding author – note that this is tougher than a REF submission). The outputs also need to be ‘new’, i.e. not used for a previous appointment or promotion within the organization.
  3. Where somebody has worked part time, or has career breaks during the period, the number of outputs are reduced pro rata.
  4. The panel assigns star ratings to each of the outputs on a similar scale to the REF (i.e. 0-4, possibly with the option of using fractional values e.g. 2.5).
  5. For appointments, candidates can then be compared on the basis of their best outputs, with appropriate pro rata adjustments. This is a much fairer system than using career totals.
  6. For promotions, a set star rating is agreed which is then uniformly applied, with appropriate pro rata for part time staff. For example, for 11 starts for Associate Professor, 13 stars for full Professor etc.

These criteria would completely replace career total judgments. Other criteria (on teaching, contributions to the university, international reputation etc.) would of course remain in place. This would be a fairer, more transparent system because universities would be recruiting and promoting people on the basis of their best research. And this system would also be aligned with REF, which would itself have benefits. Moreover, I would suggest that we use these criteria when describing biographies of senior academics. We would not say “Professor X has published over 150 papers…” Instead, we would say “Professor X’s most important work has been Y, and most important recent work has been Z”. Thus we could also eliminate ‘career total’ as a mark of esteem and instead use a ‘personal best’ esteem narrative.

Fair Recruitment, Fellowship and Promotion Panels

The second change is to ensure that all panels for recruitment, fellowship and promotion decisions are mandated to be at least one third women (and, for that matter, at least one third men). This proposal is similar to Owen Barder’s Pledge. I have sat on all male panels, and was appointed to my current job by an all male panel. I am not suggesting that all male panels are necessarily intentionally biased, but it is very easy for unintended bias to become part of such a process. This proposal would apply to recruitment panels (for all staff – including post-docs as well as permanent academic staff), fellowship decision panels and promotion panels. A one third composition could be implemented as follows:

  1. Panels of size 3-5 must contain at least 1 woman (and at least 1 man).
  2. Panels of size 5-8 must contain at least 2 women (and at least 2 men).
  3. Panels of size 9-11 must contain at least 3 women (and at least 3 men).

These compositions would help to reduce unintended bias and, in my view, improve prospects for career progression for excellent women scientists.

Goodbye to Michelle Baker

Michelle Baker left our lab today. In two months she has made considerable progress on an important project – getting to the point of having a draft of a short paper (with an interesting result). Her leaving sign-off is an exemplar of how best to leave a job (or PhD or MRes for that matter). She handed over a USB stick. On it were six self-explanatory directories:

  • Matlab code
  • Paper files (including earlier drafts)
  • Plots (all plots)
  • Poster figures (updates of figures for the poster we are making)
  • Saved Matlab results (output files)
  • Saved papers (pdfs of all research papers she cited)

Wow! I shall be showing this to everyone who will be leaving the lab in future.

We wish Michelle well and would welcome her back in any role when she returns to work.