About me

I was born in Lisbon in 1976, spent most of my childhood in Pocariça and my teens in Linda-a-Velha, and these days I'm both Portuguese and Australian. I started out in organic chemistry at FCT NOVA in 1994. That same year my uncle, António Câmara (who later won the Pessoa Prize), invited me to help at GASA, his environmental systems group at FCT NOVA, where I stayed until 1999. My job was humble: typing data from paper into databases so it could become features in virtual worlds, for projects that included a multimedia CD-ROM for Expo 98. It was written up as a book chapter, so my first publication has nothing to do with proteins. It was also where I first sat at a Silicon Graphics workstation wearing VR shutter glasses, which turned out to be handy: when I started in protein crystallography in 1998, we built models in O and TURBO-FRODO on the same kind of SGI machines.

In 1998 I started an undergraduate project in protein crystallography at ITQB NOVA in Oeiras, supervised by Pedro Matias, building the model of the split-Soret cytochrome c into the electron density. I also crystallised a seven-iron ferredoxin from a heat-loving archaeon, Acidianus ambivalens. Its crystals diffracted to 2.0 Å but were twinned, so the structure took its time: it came out in 2008. That December I spent a week at a CCP4 workshop on refinement in York. The university had run out of rooms, so Jorge Navaza, the author of the molecular replacement program AMoRe, kindly put me up in the place he was renting. From July to October 1999 came my first placement, at EMBL in Grenoble on the ESRF site, working on radiation damage in protein crystals with Raimond Ravelli, a wonderfully creative scientist who is dearly missed. From early 2000 I had a research initiation grant (BIC) in Carlos Frazão's group at ITQB, first refining a dimeric cytochrome c3 to 1.2 Å and then on the hybrid cluster protein. Ricardo Coelho and I spent a week in Juan Fontecilla-Camps's lab at the IBS in Grenoble learning to freeze crystals in liquid propane inside an anaerobic chamber, and brought the trick back to Oeiras. From September to November 2000 I was at the University of Georgia in the lab of Jean LeGall, who first described the bacterium Desulfovibrio gigas, purifying proteins in anaerobic glove boxes. Ricardo Coelho was there too, purifying proteins with me; one of those we wanted to bring back to Oeiras was a hydrogenase that stopped working if it saw any oxygen. After that, flasks of organic compounds couldn't compete with proteins and their crystals.

From October to December 2001 I went back to Grenoble for a second spell, this time at the ESRF with Ed Mitchell, who would later co-supervise my PhD. The plan was to learn phasing properly: I grew new crystals of the hybrid cluster protein without oxygen, collected data at three wavelengths around the iron edge, solved the structure by MAD and refined it to 1.3 Å, and learned XDS along the way, having only used MOSFLM and DENZO/SCALEPACK until then. Early in 2002 came HERCULES, a six-week course in Grenoble for scientists who use large facilities such as synchrotrons and neutron sources. Six weeks of lectures and experiments with PhD students from all over Europe made for close friendships: that autumn I was at a new friend's wedding in Rome, and these days we both live in the UK and are still friends. ITQB is where Larry Sieker and Ricardo Coelho taught me how to crystallise proteins, and most of what I know about crystallisation I still owe to them. My PhD (2002–2007), on an FCT scholarship, was in Maria Arménia Carrondo's lab, split between ITQB and the ESRF, with Carlos Frazão and Ed Mitchell as my supervisors. Grenoble is where I fell for synchrotrons, and I've never quite got over it.

In December 2002 I'd done an EMBO practical course on protein expression, purification and crystallisation (PEPC3) at EMBL Hamburg, and it came in handy. Around the end of 2003, halfway through the PhD, I took on a second project: Maria Arménia set up a collaboration with Isabel Sá-Correia's group at IST in Lisbon. Our lab had always worked on metalloproteins, purified the way Jean LeGall taught: grow huge cultures and extract the proteins the bacteria make in abundance. So I became the first in the lab to overexpress a protein and to make it with selenomethionine. The IST team knew overexpression but not selenomethionine, so we learned together, and my thesis was better for having both proteins in it.

The new protein was UgpG, a glucose-1-phosphate uridylyltransferase, and by early August 2005 I had seven crystal forms and no structure. Molecular replacement wasn't getting anywhere, and the selenomethionine data had 88 selenium sites to find (11 in each of the 8 molecules), which was no small job. In the early hours of the last day of beam before the ESRF's August shutdown, an osmium derivative started showing a lovely anomalous signal, but the machine was due to be switched off before I could finish. Ed told me straight: there wasn't enough time, and if I wanted the structure before the shutdown I'd have to ring the control room and ask them to keep the machine running. Was that even possible? He didn't know, and he wouldn't make the call for me: my decision, my phone call. With shaking hands and a wobbly voice, I rang. They said yes. The whole ESRF kept running for another 25 minutes because of me, and that was enough for a complete, highly redundant dataset and a solved structure. The selenium sites earned their keep later, pinning down the sequence in maps that weren't easy to build into. I spent the rest of the summer building, and by the end of September I had a complete, refined structure (paper).

Next came membrane proteins with Martin Caffrey: first at the University of Limerick, as a postdoc on his US National Institutes of Health grants, and then at Trinity College Dublin on my own Marie Curie fellowship (the first grant I wrote, and the first I won), working on ABC transporters from Pseudomonas aeruginosa. Limerick was a leap. The lab had no crystallography of its own yet (until then it had been done with outside collaborators), so my first jobs were booking synchrotron beamtime and buying the loops, dewars and the rest of the kit we needed to collect data at all. As the resident crystallographer I was involved in every project, which made for exciting, rewarding and very busy years: between 2008 and 2011 I ran 31 trips to synchrotrons, about one a month, most of them to the Advanced Photon Source near Chicago, and others to the Swiss Light Source, the ESRF and Diamond.

After a few trips on my own, especially the 48-hour blocks of beamtime the APS gave us every month, I realised I needed help, so I started bringing students from the lab along, sometimes before they had crystals of their own. That's where I found out how rewarding it is to help other people grow, as scientists and as people. My imposter syndrome was loud: what if we hit a crystallographic problem I didn't know how to solve? We grew crystals inside the lipidic cubic phase, a lipid gel with roughly the texture of toothpaste, and some of that work ended up in Nature, including a receptor structure with Brian Kobilka's group. I'd love to tell you it was all part of a grand plan. It wasn't. It was good mentors, great colleagues and a lot of crystals that didn't grow. And plenty of willpower and persistence too, for example in learning to get crystals out of the glass sandwiches we grew them in.

In 2011 I swapped research for beamline science at the Australian Synchrotron in Melbourne, where Tom Caradoc-Davies hired me and was my manager and mentor for the next eight years. He had a lot of patience with me, kept encouraging me to chase what I wanted to do, and turned a keen but fairly ordinary academic into an instrument scientist.

The doubts still came back, in a new form: suddenly I was the person who was meant to know why the robot had stopped, the detector had gone quiet or the data looked strange. My cure has always been the same: learn the thing properly, write it down, and share it. I wrote elog entries and mini how-tos about everything, rather selfishly at first, even for problems other people had fixed, because I thought I might need them one day. Over two or three years that habit made me the go-to troubleshooter, and the notes became one of the best sources of know-how on the beamlines. By 2018 I'd written more than 4,000 entries in the facility logbook, the second most of anyone there. Outside the beamline, seven years on the SCANZ executive taught me how much a small society can do with good volunteers (and a decent membership database). I also helped prepare the winning bid to bring the 2023 IUCr World Congress to Melbourne, and then joined its organising committee. Python courses, two Red Hat certifications and many patient colleagues later, it still works.

In 2019 I moved to Diamond in Oxfordshire, just in time for a pandemic. During the lockdowns I was one of the key workers keeping I04 running for the COVID Moonshot.

Since then, much of my work on I04 has been about quality, and about making good data easier to get. I brought the checklist habit with me: a software beamline setup checklist, now MX Beamline Setup, so every start-up is checked the same way and nothing depends on someone's memory. I led the move to dose-aware and then dose-driven data collection: first making users aware of the X-ray dose their crystals were getting, then letting the beamline choose exposures from a target dose (2025 paper). At IUCr 2026 I set out the wider framework for teaching and collecting data this way in a poster, Strategic dose allocation. In 2024 and 2025 I led a BBSRC-funded project that brought machine-learning sample centring to I04. I also teach data collection at CCP4 workshops every year since 2019, some years at two different schools, and every year on Diamond's own MX user training, which is where all those old notes keep paying off. In 2026 that training was split into three courses for different levels of experience; I taught on the introduction for new users in February. For Diamond's 2026 Beyond Automation course I came up with four hands-on case studies for I04, from detector count-rate problems that merging statistics hide to spreading the dose along needle crystals. I ran two of them and colleagues ran the other two, and I drew each one as a comic strip (with some AI help).

These days I also co-supervise PhD students (at the moment Paige Taylor, who has just moved from Australia to Diamond) and train new colleagues to become instrument scientists, trying to pass on some of the patience so many people showed me. I'm also a mental health first aider at work, sit on Didcot Town Council, and still get a small thrill every time a new structure appears out of the data. Enough of a thrill that I wrote a display for the screen in front of I04, which shows a rotating loop of newly released structures collected on the beamline. It's an idea I first tried in Melbourne as the "MX TV", on old screens rescued from recycling.

I'm particularly interested in making MX beamlines more automated without compromising the quality of the science or the safety of the equipment, so that robots and goniometers don't collide and the direct beam never reaches a detector. That includes better ways to collect data with modern pixel-array detectors, collecting complete datasets from several wedges on the same or different crystals, the challenges of small and very small X-ray beams (beam stability, a small sphere of confusion and so on), and better ways to process the raw data from all these setups.

I also care about the user's experience of a beamline, whether they're on site or working remotely from their own lab: how experiments are designed, how data collection software presents itself, and how raw data and metadata are archived and shared so the next experiment can learn from the last. Modern beams and detectors produce an overwhelming amount of data, and that calls for fresh thinking rather than doing things the same way again and again. Behind all of that sits the computing, and I still enjoy that side too: Linux, virtual machines, containers and open-source tools that keep a beamline reliable and easy to look after.

In structural biology, I'm drawn to the interfaces of cells and organelles: how bacteria import what they need to survive, how energy is turned into proton gradients across membranes and those gradients into ATP, how viral capsids attach to cells and assemble inside them, and how cargo proteins move molecules into and out of the nucleus, mitochondria and Golgi apparatus.

Just as much, I'm interested in the people who make a facility work: training users and new staff, mentoring early-career scientists, and helping colleagues grow into new roles. Wellbeing matters too; I'm a mental health first aider at work.

I also love talking about science with people who don't do it for a living. Open days energise me like little else, I've given a general-audience talk at the Wantage Café Scientifique, and since 2023 I've joined school career carousels (a sort of speed dating, where teenagers quiz me about my job) and run mock job interviews with students.