The white tigers of Rewa and Clifton: Mendelism and Bristol Zoo

Posted on by sian.c.harris

On 20-21 July we will be welcoming people to Bristol and online around the world to our Mendel at 200 conference. For those who will be joining us face-to-face, there is an opportunity to visit the historic Bristol Zoo Gardens. George Davey Smith shares the story of some of the early zoo residents and how they relate to Mendel’s discoveries.

The 20th July is the bicentenary of the birth of Gregor Mendel – the so-called “father of genetics”. We are marking the occasion by holding a two-day (20th-21st July) meeting at Bristol Zoo – with a free online attendance option, with a stellar cast of internationally recognised experts in the history, ethical discussions and latest research building on Mendel’s work.

The meeting is being held at Bristol Zoo, the world’s oldest provincial zoo, which opened in 1836, when Gregor Mendel was only 14. It provides a glorious setting for such a meeting, and also played a part in an international Mendelian drama.

In 1963 the zoo acquired two white tigers – Champak and Chameli – bought from the Maharaja of Rewa for more than £100,000 in today’s money. These extraordinary animals – the only ones of their type in Europe – increased the attendance figures, and by 1970 the zoo held one third of the captive world white tiger population. At this time each tiger was valued at around £200,000 in today’s money. But where had the tigers come from, and what is the link with Mendelism?

The Bristol Zoo white tigers came from a lineage that was established in captivity by the Maharaja of Rewa, then a princely state – and now part of Madhya Pradesh – in India. They had been seen in the wild around Rewa for half a century, but were not true albino tigers, having brownish stripes on an off-white background and blue eyes (true albinos lack any pigment, have pink eyes and are snowy white all over). The Maharaja named the white male tiger cub he first obtained Mohan. He kept Mohan with an orange (i.e. normally-coloured) tigress, Begum, who, in three litters, gave birth to 10 normally-coloured cubs. Mohan mated with one of his daughters, Radha, resulting in 14 offspring, some white and some orange.

Mendel famously crossed peas with different features – including having yellow or green seeds – to demonstrate the regularities in heredity that now bear his name. Using contemporary terminology and interpretation, white tigers can be used to illustrate these regularities. Mohan must have received a copy each of the same allele (labelled w in the figure below) from his mum and his dad. At each location in the genome, individuals have two alleles. When germ cells (sperm and ova) are formed, it is a matter of chance which allele ends up in each germ cell. Begum had two orange (W) alleles; she and Mohan can be referred to as homozygous for the coat colour trait. All of their offspring would thus have a w from Mohan and a W from Begum, be designated Ww and referred to as heterozygous at this locus. W is the dominant allele, thus all the offspring were orange coloured, as in the first row of the figure below. When Mohan mated with his daughter Radha (second row of the figure), each offspring will have received one allele by chance from both parents, and (with large numbers of offspring) the ratio of colours would be 1 orange to 1 white. When two orange heterozygous Ww tigers are mated – such as two of the grandchildren of Mohan and Begum, Ramana and Ramani, were – their offspring will be in a 1-2-1 genetic ratio (WW, 2Ww, ww), and a 3 to 1 orange to white colour ratio.

 

(Image source)

As with Mendel’s extensive pea crosses, the theoretical expectations based on a model that the white colour is recessive to orange and that alleles segregate equally can be tested against data. As the table below shows, this expectation is close to what was observed.

Indeed, more recently the causal genetic variant has been identified as one leading to a single amino acid change in the transporter protein SLC45A2 (although, as with everything in genetics, there is greater complexity hidden in the spectrum of coat colour variation within white tigers).

It is clear there was considerable interest in both the genetics of the captive white tigers and in their financial value. When Champak (who was male) and Chameli (who was female) arrived in Bristol in 1963 thoughts turned quickly to their potential offspring. Chameli produced three litters – all white, of course – of 14 cubs in all. All four of the first litter died soon after birth and one of the second litter also died young and was devoured by Chameli. The third litter (see the photograph below, with growth monitoring perhaps being illustrated) also prefigured an unhappy story, of generally short lives. Indeed, one of the five was already dead by the time this early photograph was taken. Something was clearly amiss, and it was not anything specific to Bristol.

 

The reason for the poor health of the white tigers is not to do with the single amino acid change, but rather with the extreme inbreeding which was employed to extend the highly profitable white tiger lineage. The first cubs – including Champak and Chameli in Bristol – were a result of a father-daughter mating, and were siblings who were subsequently mated. Ramana and Ramani (mentioned above) were also both grandchildren of the same grandparents, whilst also being siblings. This inbreeding led to high mortality and congenital facial, eye, gastrointestinal tract, cardiac, kidney and other conditions. This was recognised early, and in the Bristol Zoo genealogy (below) the reason Seeta was exchanged for the white tiger Roop from the Delhi Zoological Park (where I witnessed many white tigers when living in the city in 2009 and 2010) was both to generate a better sex balance and in an effort to reduce the extent of inbreeding.

There has also been the suggestion that inbreeding has led to white tigers being more aggressive (perhaps from perceptual difficulties and stemming from fear). High-profile stories, such as when the white tiger Jupiter killed Chuck Lizza and Joy Holliday of “Cat Dancers” (an exotic tiger entertainment act), and the even more sensational story of the white tiger Montecore (which apparently translates into “maneater”) nearly killing Roy Horn of Las Vegas megastars Siegfried and Roy certainly raised the profile of this possibility (although in the latter case it is difficult to believe anything about the story). However the apparently innocent play below might be less benign than it looks.

As the news of the poor health due to inbreeding suffered by the white tigers became more widely known, their visitor attraction fell, and by mid-1985 they were gone from the Bristol Zoo collection. The inbreeding and display of the white tigers would certainly not fit with the ethos of the zoo (advanced in 2008) to “maintain and defend biodiversity through breeding endangered species, conserving threatened species and habitats and promoting a wider understanding of the natural world”.

Sadly, the Bristol Zoo Gardens site will be closing its doors to the public at the end of this summer. All the animals and activities are moving to the much-bigger Wild Place site in South Gloucestershire. This newer site has been purpose built with the zoo’s conservation goals at its heart, but it will be sad to say goodbye to the zoo that has been in the heart of Clifton in Bristol for nearly two centuries. Those of us who will be attending the conference in person will have a chance to visit the zoo site as it starts its ‘Big Summer Shutdown’ (admission to the zoo is part of the package for all face-to-face delegates at our conference).

We will also be joined by participants and speakers from all over the world via zoom. Registration for the conference, either face-to-face or online, is available up to the day (although in-person places are limited). Find out more information and register here.

Note: Andrew Flack’s excellent book on Bristol Zoo, “The Wild Within: Histories of a Landmark British Zoo. 2018, Charlottesville: University of Virginia Press” alerted me to the story of the white tigers in Bristol Zoo.

From a love of puzzles to studies on BMI – what Mendel’s legacy means to me, and to my cat

Posted on by sian.c.harris

In the first of a series of blog posts celebrating 200 years since the birth of Gregor Mendel, Lavinia Paternoster shares how learning about genetics at school shaped her future career – and introduces us to a cat called Mendel

 

A cat called Mendel.
A cat called Mendel

For as long as I can remember I’ve loved spotting patterns, spending hours as a child playing logic puzzles and, more recently, Sudoku. I love how just a few simple rules can be applied to break the code of seemingly complex patterns. So when I was introduced to Mendel’s pea experiments during my A-levels it was like I got to use my nerdy love of puzzles in the classroom. Compared to how hard I found languages and chemistry, I couldn’t believe that solving these little crosses to determine the genetic inheritance of pea traits counted as work. I had a very supportive biology teacher who nurtured my passion by sending me home with a jar of fruit flies over the Easter holidays to perform my own inheritance crosses (more in homage to Morgan’s drosophila crosses, but quicker and requiring less horticultural skills than growing pea plants). I was hooked and quickly signed up to study genetics at university. 

Today I still love the simplicity in the way that the laws of genetic inheritance work to influence even the most complex of human traits. Now working on human traits such as eczema, BMI and even how a disease progresses over time, most of my work involves the simplest of statistical tests (performed millions of times, in an approach called genome-wide association  studies) to identify which variants in our genomes influence these important outcomes. 

I often think about my earliest introduction to genetic inheritance and how lucky I was to find my imagination captured by those beautifully simple genetic crosses performed by Mendel. Naming my own cat in his honour, I often find myself chatting to a random passers by outside our house about Mendel and his pea experiments. Whilst glad I share some of Mendel’s (the man not the cat) love of genetic inheritance, I definitely do not also share his talent in the greenhouse, struggling to keep the most low maintenance of plants alive. But I somewhat blame Mendel’s love of digging (the cat, not the man, this time)!

 

  • To celebrate Mendel’s 200th birthday we are holding a two-day conference, online and in-person in Bristol on 20-21 July. For more information and to sign up, see our Mendel at 200 pages and follow #Mendel200 on social media for Mendel activities around the world.

 

Why siblings are interesting for genome-wide association studies

Posted on by sian.c.harris

Neil Davies discusses a new paper on a genome-wide association study of almost 180,000 siblings and discusses what additional insight siblings bring to such studies.

Thousands of genome-wide association studies (GWAS) have been published, however, the vast majority have used samples of unrelated individuals. We have recently published a sibling GWAS published in Nature Genetics. In our study, we used almost 180,000 siblings across 19 studies from around the world. But why are siblings interesting for GWAS?

GWAS have already identified tens of thousands of single nucleotide polymorphisms (SNPs) related to phenotypes – using samples of unrelated individuals. However, correlation is not equal to causation. Increasing evidence suggests these associations can be driven by more than individual-level biological effects.

There can be three key sources of bias. The first potential bias is population stratification. This means the differences in the frequency of the genetic variants that relate to phenotypic differences. For example, Iron Brew consumption will associate with variants more common in Scotland. These associations are biased evidence of the causal effect of the variant on the phenotype!

The second bias is assortative mating. People don’t mate at random. For example, studies have shown that more educated people tend to have more educated and taller partners. Such trends can result in biased associations between SNPs and phenotypes in the offspring.

The third bias is indirect parental genetic effects (also known as dynastic effects).

In these, the genotype is expressed in parents, which in turn affects offspring outcomes. One example of this is that the education of parents may influence educational outcomes in the offspring, again biasing SNP-phenotype associations.

How can data from siblings help overcome these biases? Siblings inherit their genetic variants from their parents at random. They are nature’s randomized control trials. If the siblings who share the genotype have more similar trait measures, researchers can be more confident that the genotype is influencing the trait directly.

Looking at the differences between siblings controls for each of the sources of bias above.

Which phenotypes suffer most from these biases? In our Nature Genetics paper, we estimated the shrinkage from the population to sibling estimates for 25 phenotypes, to see which suffered most from these biases. We estimated this by looking at how much the associations shrunk between the population estimates (without comparing within siblings), to the within sibling estimates. The larger shrinkage in the LD-score regression plot below indicates more bias.

We found that previously reported genome-wide association study (GWAS) associations, which typically use more widely available population samples of unrelated individuals, tend to overestimate direct effects for many traits including educational attainment, cognitive ability, age when first gave birth, whether someone has ever smoked, depressive symptoms and number of children. We also found that estimates of heritability, genetic correlations and other genetic analysis methods could substantially differ when calculated using estimates from siblings.

Biases do affect genetic correlations

A major finding from our research was that these biases do affect genetic correlations. When we use sibling cohorts, the genetic correlations from LD-score regression between educational attainment and traits such as height and BMI are not detected. Note the change in power and precision in the plot below. This suggests that the correlations that are detected in population samples are unlikely to be due to a causal effect of the genetic variants in the individuals.

Are recent findings on polygenic adaption robust to these biases? Yes, height is likely to be under polygenic selection. This suggests that selective pressures in the human population have affected the number of height-associated alleles in the population. This could lead to changes in the average height of the population over multiple generations.

Are sibling samples “better” than “population” samples?

Whether sibling samples such as we use in our study are “better” than population studies depends on the question you want to look at. Large population-based samples of unrelated individuals are great if you want to discover new genetic variants associated with a disease or other outcomes, or are interested purely in prediction.

However, if you are interested in understanding why genetic variants associated with an outcome like height, BMI, or education, then family studies can provide a powerful source of evidence. In this paper, we only looked at a very small number of phenotypes, but these results suggest that these biases are more likely for social/behavioural phenotypes, and more biological ones are less likely to be biased.

What’s next? The international collaboration established for this study is continuing to work together and explore these issues further. The next steps include using bigger samples of siblings and estimating the relative contribution of these sources of bias using samples of parent-offspring trios.

A massive thanks to all our co-authors – an international group of 100 scientists were involved in this study – and many, many others. Amazing being able to work with you all!

Read the paper

Read the press release

Read our FAQs

Webinar on 19th May honours the career of “Aspirin Man”

Posted on by sian.c.harris

In the 1970s two randomised trials of aspirin led by Professor Peter Elwood from the MRC Epidemiology Unit, South Wales made the headlines for finding that a low dose of aspirin had beneficial effects for patients who had had a heart attack.

This was just one of many important discoveries from over 50 years of epidemiological research carried out in South Wales in MRC units, including the Epidemiology Unit initially directed by Archie Cochrane, and then by Peter Elwood.

Peter joined the unit in South Wales in 1963 and led it from 1974 until it closed in 1995. Over that time he led on converting the unit from one which had researched respiratory disease and other issues to one with a focus on cardiovascular disease (which had shown increasing rates since the 2nd world war, whilst pneumoconiosis and tuberculosis decreased). Since the unit closed in 1995 he has continued working, producing more than many people do who are still employed.

In recognition of Peter’s long and valuable career, IEU’s Professor George Davey-Smith and Professor John Gallagher (Director of the Dementia Platform UK, University of Oxford), who both worked with Peter in the unit, are organising a half-day meeting in Peter’s honour on 19th May, 14:00-17:30 BST, in Oxford and online.

The meeting will feature speakers from Bristol, Oxford and UCL including Professors Nishi Chaturvedi, Andy Ness, Sir Michael Marmot and Sir Richard Peto, as well as Peter himself, discussing important topics in epidemiology. These include the role of alcohol in cardiovascular disease; how diet influences disease risk; potential causal relationships between diabetes and dementia; health inequalities, productive research environments and aspirin.

See the full programme.

Register for the online event (free) on the EventBrite page.

Biography of Professor Peter Elwood

More about the history of epidemiology in South Wales.

Erasing the stain: Challenging the stigma of opioid substitution treatment. Findings from a stakeholder workshop

Posted on by lindsey.pike

Author: Vicky Carlisle. Twitter: @Vic_Carlisle, Email: vicky.carlisle@bristol.ac.uk

On Wednesday 7th July 2021, I brought together key stakeholders with an interest in improving opioid substitution treatment (OST) from across the UK. This included people with lived experience, Public Health England staff, local authority public health practitioners, treatment service leads, pharmacists and academics. We discussed the findings of my recently completed PhD, and together we considered the next stages of developing an intervention to improve OST.

A summary of my research

For those not familiar with the topic, OST refers to the treatment of opioid dependency with either methadone or buprenorphine (alongside psychosocial support). Through my research, I wanted to understand what the key facilitators and barriers are to people ‘recovering’ in OST. To do this, I drew on both quantitative and qualitative methodologies. I found that loneliness, isolation and experiences of trauma and stigma were key barriers to recovery; whereas positive social support, discovering a sense of purpose and continuity of care were valuable facilitators.

Importantly, some factors appear to act as both facilitators and barriers to recovery in OST. For instance, I found that some service users used isolation as a form of self-protection (to shield themselves from negative influences), however this often left them feeling lonely and disconnected from the potential benefits offered by developing more positive social support networks.

Undoubtedly, the strongest barrier to recovery was stigma. Service users told me that they experience stigma from a range of sources, including from family and friends, healthcare professionals and members of the wider community. I found similar patterns in the literature review that I carried out (Carlisle et al, 2020). Stigma is like a stain where an individuals’ entire identity is defined by a single, negative attribute. In the case of OST, individuals may possess overlapping stigmatised identities of ‘OST service user’, ‘drug user’ and ‘injecting drug user’. Some will be further stigmatised due to experiencing homelessness, being HIV or Hepatitis C positive or through involvement in sex-work.

“I found that loneliness, isolation and experiences of trauma and stigma were key barriers to recovery”

Community pharmacies are one environment where service users report experiencing a great deal of stigma. Unlike customers collecting other prescriptions, many OST service users receive their medications (methadone/buprenorphine) through an arrangement known as ‘supervised consumption’. This means they must be observed taking their medication by a pharmacist to ensure that it is not diverted to others. This is often conducted in full view of other customers, despite guidelines which recommend that this takes place in a private room or screened area. This leaves OST service users open to the scrutiny of the ‘public gaze’.

My findings have several implications in relation to stigma. Firstly, OST service users receive poorer care than other members of society in healthcare settings, which may result in them avoiding seeking help from drug treatment and for other health conditions. Secondly, stigmatising OST service users makes community re-integration extremely challenging and this has been directly linked to individuals returning to drug using networks as it is somewhere they feel a sense of belonging. The ultimate impact of being repeatedly exposed to stigma is an internalisation of these judgements, resulting in feelings of shame and worthlessness – again impacting on individuals’ ability to seek help and develop supportive new relationships with others.

Figure 1: Key facilitators and barriers to recovery, retention and completion in OST at each level of the socioecological model. Stigma is present at every level of the system.

What we discussed during the workshop

Being able to present these findings to key stakeholders was a real highlight of my PhD work; it’s not often that you have the ear of so many invested and engaged individuals in one ‘room’ (albeit a Zoom room!). The findings of my PhD chimed closely with the experiences of those in the room and would be further reflected the next day when Dame Carol Black’s Review of Drugs Part 2 was published, which made specific reference to stigma.

After I presented a short overview of my PhD findings, attendees spent time in small groups discussing how we might address OST stigma at each level of the socioecological system (see figure 1, above). A common thread that ran through each of the groups’ discussions was the importance of embedding interventions within trauma-informed frameworks. Attendees felt that increasing others’ understanding of the impact of trauma and ‘adverse childhood experiences’ (ACEs) may be a key mechanism by which to reduce stigma towards OST service users.

Indeed, a recent study found promising results in relation to this – that increasing the public’s awareness of the role of ACEs in substance use reduced stigmatising attitudes towards people who use drugs (Sumnall et al, 2021). Workshop attendees suggested that this outcome could be achieved through trauma-informed training of all individuals who might work with OST service users, such as pharmacists, the police and medical professionals, as well as those who work in healthcare settings, such as receptionists.

At the individual level there was a discussion about the way that stigma trickles down the socioecological system, resulting in self-stigma or internalised stigma. People felt that the best way to reduce this was to tackle stigma higher upstream first.

When thinking about reducing stigma in everyday inter-personal interactions, people highlighted the importance of using non-stigmatising language. For those who are interested (and I think we all should be!) the Scottish Drug Forum has published an excellent guide here.

Some excellent suggestions were made for reducing stigma that individuals experience in organisations such as pharmacies, hospitals and other settings. This is something that Dr Jenny Scott and I discussed in a recent article for the Pharmaceutical Journal (Scott & Carlisle, 2021). One attendee suggested the introduction of positive role-models within organisations who could be an exemplar of positive behaviour for others (a ‘stigma champion’ perhaps?). Training was identified as a key mechanism through which stigma could be reduced in organisations, including through exposure to people who use drugs (PWUD) and OST service users during training programmes. It was stressed however, that this should be carefully managed to ensure that a range of voices are presented and not just ones supporting dominant discourses around abstinence-based recovery.

Suggestions for improving community integration included increasing access to volunteering opportunities – something that people felt has been impacted by reduced funding to recovery services in recent years. It was also suggested that community and faith leaders could be a potential target for education around reducing stigma and understanding the impact of trauma, as these individuals may be best placed to have conversations about stigma with members of their communities.

Finally, there were some thoughtful discussions around the best way to influence policy to reduce stigma. The importance of showing policymakers the evidence-base from previous successful strategies was highlighted. Something that resulted in a lively debate was the issue of supervised consumption with arguments both for and against (this is also relevant at the organisational level). The group summarised that whilst diversion of medications was a risk for some, a blanket approach to supervised consumption and/or daily collections exposes individuals to stigma in the pharmacy, which leaves individuals vulnerable to dropping out of treatment. It was pointed out that supervised consumption policies were quickly relaxed at the start of Covid-19 restrictions – something that appears to have been done safely and with benefits to service users. It was also highlighted that supervised consumption in OST is inherently stigmatising, as users of other addictive drugs with overdose potential, such as other prescribed opioids and benzodiazepines, are not subjected to the same regulations. This sends a clear message to OST service users that they cannot be trusted. Other key suggestions were:

  • Communicating with CQCs and Royal Colleges, who may be particularly interested in understanding how people are treated in their services.
  • Drawing on existing stigma policies from other arenas e.g. mental health.
  • Highlighting the fiscal benefits of reducing stigma to key decision makers.
  • Tapping into plans for the new Police and Crime Commissioners, who have a trauma sub-group.
  • Linking into work with ADDER areas, which may provide the evidence for ‘what works’.

What next?

I am now planning to apply for further funding to develop an intervention to reduce organisational stigma towards OST service users. The involvement of service users and other key stakeholders will be crucial in every step of that process, so I will be putting together a steering group as well as seeking out collaborations with academics internationally that have expertise and an interest in this area. I was really pleased to see that Dame Carol Black’s second report makes some concrete recommendations around reducing stigma towards people who use drugs. I hope therefore to be able to work with the current momentum to make OST safer and more attractive to those whose lives depend on it.

I’d like to extend my gratitude to all of the attendees at the workshop and to Bristol’s Drug and Alcohol Health Integration Team (HIT) for supporting this event. If you are an individual with lived experience of OST, an academic, or any other stakeholder working in this area and would like to be involved with future developments, please get in touch with me at vicky.carlisle@bristol.ac.uk or find me on Twitter at @Vic_Carlisle.

References

Carlisle, V., Maynard, O., Padmanathan, P., Hickman, M., Thomas, K. H., & Kesten, J. (2020, September 7). Factors influencing recovery in opioid substitution treatment: a systematic review and thematic synthesis. https://doi.org/10.31234/osf.io/f6c3p

Scott, J & Carlisle, V (2021). A pharmacy resolution for 2021: let’s improve the way patients with addiction are treated. The Pharmaceutical Journal. https://pharmaceutical-journal.com/article/opinion/a-pharmacy-resolution-for-2021-lets-improve-the-way-patients-with-addiction-are-treated

Sumnall, H. R., Hamilton, I., Atkinson, A. M., Montgomery, C., & Gage, S. H. (2021). Representation of adverse childhood experiences is associated with lower public stigma towards people who use drugs: an exploratory experimental study. Drugs: Education, Prevention and Policy, 28(3), 227-239. https://doi.org/10.1080/09687637.2020.1820450

This blog was originally posted on the TARG blog on the 1 October 2021.

COVID-19: Pandemics and ‘Infodemics’

Posted on by lindsey.pike

JGI Seed Corn funded project 

Drs Luisa Zuccolo and Cheryl McQuire, Department of Population Health Sciences, Bristol Medical School, University of Bristol. 

The problem 

Soon after the World Health Organisation (WHO) declared COVID-19 a pandemic on March 11th 2020, the UN declared the start of an infodemic, highlighting the danger posed by the fast spreading of unchecked misinformation. Defined as an overabundance of information, including deliberate efforts to disseminate incorrect information, the COVID-19 infodemic has exacerbated public mistrust and jeopardised public health.  

Social media platforms remain a leading contributor to the rapid spread of COVID-19 misinformation. Despite urgent calls from the WHO to combat this, public health responses have been severely limited. In this project, we took steps to begin to understand and address this problem.  

We believe that it is imperative that public health researchers evolve and develop the skills and collaborations necessary to combat misinformation in the social media landscape. For this reason, in Autumn 2020 we extended our interest in public health messaging, usually around promoting healthy behaviours during pregnancy, to study COVID-19 misinformation on social media. 

We wanted to know:  

What is the nature, extent and reach of misinformation about face masks on Twitter during the COVID-19 pandemic? 

To answer this question we aimed to: 

  1. Upskill public health researchers in the data capture and analysis methods required for social media data research; 
  2. Work collaboratively with Research IT and Research Software Engineer colleagues to conduct a pilot study harnessing social media data to explore misinformation. 

The team 

Dr Cheryl McQuire got the project funded and off the ground. Dr Luisa Zuccolo led it through to completion. Dr Maria Sobczyk checked the data and analysed our preliminary data. Research IT colleagues, led by Mr Mike Jones, helped to develop the search strategy and built a data pipeline to retrieve and store Twitter data using customised application programming interfaces (APIs) accessed through an academic Twitter account. Research Software Engineering colleagues, led by Dr Christopher Woods, provided consultancy services and advised on the analysis plan and technical execution of the project. 

Cheryl McQuire, Luisa Zuccolo, Maria Sobcyzk, Mike Jones, Christopher Woods. (Left to Right)

Too much information?!

Initial testing of the Twitter API showed that keywords, such as ‘mask’ and ‘masks’, returned an unmanageable amount of data, and our queries would often crash due to an overload of Twitter servers (503-type errors). To address this, we sought to reduce the number of results, while maintaining a broad coverage of the first year of the pandemic (March 2020-April 2021).

Specifically, we:

I) Searched for hashtags rather than keywords, restricting to English language.

II) Requested original tweets only, omitting replies and retweets.

III)  Broke each month down into its individual days in our search queries to minimise the risk of overload.

IV) Developed Python scripts to query the Twitter API and process the results into a series of CSV files containing anonymised tweets, metadata and metrics about the tweets (no. of likes, retweets etc.), and details and metrics about the author (no. of followers etc.).

V) Merged data into a single CSV file with all the tweets for each calendar month after removing duplicates.

What did we find?

Our search strategy delivered over three million tweets. Just under half of these were filtered out by removing commercial URLs and undesired keywords, the remaining 1.7m tweets by ~700k users were analysed using standard and customized R scripts.

First, we used unsupervised methods to describe any and all Twitter activity picked up by our broad searches (whether classified as misinformation or not). The timeline of this activity revealed clear peaks around the UK-enforced mask mandates in June and September 2020.

We further described the entire corpus of tweets on face masks by mapping the network of its most common bigrams and performing sentiment analysis.

 

 

 

We then quantified the nature and extent of misinformation through topic modelling, and used simple counts of likes to estimate the reach of misinformation. We used semi-supervised methods including manual keyword searches to look for established types of misinformation such as face masks restricting oxygen supply. These revealed that the risk of bacterial/fungal infection was the most common type of misinformation, followed by restriction of oxygen supply, although the extent of misinformation on the risks of infection decreased as the pandemic unfolded.

Extent of misinformation (no tweets), according to its nature: 1- gas exchange/oxygen deprivation, 2- risk of bacterial/fungal infection, 3- ineffectiveness in reducing transmission, 4- poor learning outcomes in schools.

 

Relative to the volume of tweets including the hashtags relevant to face masks (~1.7m), our searches uncovered less than 3.5% unique tweets containing one of the four types of misinformation against mask usage.

A summary of the nature, extent and reach of misinformation on face masks on Twitter – results from manual keywords search (semi-supervised topic modelling).

A more in-depth analysis of the results attributed to the 4 main misinformation topics by the semi-supervised method revealed a number of potentially spurious topics. Refinements of these methods including iterative fine-tuning were beyond the scope of this pilot analysis.

 

Our initial exploration of Twitter data for public health messaging also revealed common pitfalls of mining Twitter data, including the need for a selective search strategy when using academic Twitter accounts, hashtag ‘hijacking’ meaning most tweets were irrelevant, imperfect Twitter language filters and ads often exploiting user mentions.

Next steps

We hope to secure further funding to follow-up this pilot project. By expanding our collaboration network, we aim to improve the way we tackle misinformation in the public health domain, ultimately increasing the impact of this work. If you’re interested in health messaging, misinformation and social media, we would love to hear from you – @Luisa_Zu and @cheryl_mcquire.

Note:

This blog post was original written for the Jean Golding Institute blog

Metadata errors: an under-appreciated source of bias in Mendelian randomization studies

Posted on by lindsey.pike

Dr Philip Haycock

In two-sample Mendelian randomization (MR), a type of epidemiological method, we combine the results from different genetic studies to study the causal relationship between human characteristics and disease. For example, we might take results from a genetic study of smoking and a different genetic study of cancer. We can combine their results to understand whether smoking might be a cause of cancer. If the same position in the genome is associated with smoking in the first study and with cancer in the other study, this can provide evidence that smoking is a causal factor in cancer. However, it’s also possible that this position in the genome could be related to smoking and cancer via separate pathways. This phenomenon is known as “horizontal pleiotropy” and is a common source of bias in Mendelian randomization research. 

Another, often under-appreciated, source of bias are errors in metadata. To understand this we need to understand what genetic results look like in practice. Below is an example of a genetic results file with 5 rows and 6 columns (a typical file might actually have several million rows).

Example of genetic results

Each row refers to a single position in the human genome that varies between people. These positions are referred to as “genetic variants” (also known as polymorphisms). The particular type of variant that an individual carries is known as their allele. 

Below is an example of metadata. The metadata helps us understand the contents of the results file. It tells us what the columns represent. 

Example of metadata

Some columns in the results file will describe the relationship (“effect” column) between the genetic variant and some human characteristic (e.g. smoking) and there will be additional columns that help researchers interpret this relationship. These additional columns include things like the identity of the allele that is used to model the relationship (e.g. if people have allele “A” they may be more likely to smoke compared to people without this allele) or information on how common the allele is in the population. These columns are also known as the “effect allele” and “effect allele frequency” columns. Metadata errors refer to mistakes in how these columns are reported. For example, maybe allele1 is reported as the effect allele column when in fact it should have been allele 2 that is described in this way. Sometimes the information provided in metadata is ambiguous. For example, the metadata tells us that the “freq” column represents allele frequency but there are two alleles. Is this the frequency of allele1 or allele2? We can’t be sure. Another type of error refers to mistakes in the reported results, for example reporting that a genetic variant increases the probability that a person smokes when in reality it has no effect (in other words the effect is zero). This is known as a summary data error. Failure to identify these errors can lead to mistakes in Mendelian randomization analyses, such as finding that smoking protects against cancer (when we know the opposite is true). 

As research complexity increases, so does the potential for errors

These types of errors were fairly easy to avoid during the early years of Mendelian randomization research, when studies tended to be hypothesis-driven and focused on small numbers of relationships (although errors still occurred). Mendelian randomization study designs are, however, increasingly complex and hypothesis-free, sometimes assessing relationships amongst 100s or even 1000s of characteristics and diseases. New online platforms and databases that collate genetic results from many different sources, and provide tools that can automate analyses, make these studies easier to undertake than ever before. The downside is that they probably make meta and summary data errors more likely. 

Maximising metadata quality to reduce errors

We address this issue in a new pre-print: “Design and quality control of large scale Mendelian randomization studies”. We present an R package and set of quality control tools that identify meta and summary data errors, which we developed for the Fatty Acids in Cancer Mendelian Randomization Collaboration (FAMRC). The FAMRC is a pan-cancer MR study that seeks to evaluate the causal relevance of fatty acids for risk of major cancers. We wanted to maximise the quality of the genetic study results we collected from the cancer studies, to ensure the integrity of our Mendelian randomization analyses. After implementing our tools, we found major meta and summary data errors in 7 (13%) of 55 genetic studies in the FAMRC.

What types of metadata errors did we find?

The basic principle of our quality control approach is to identify errors through 

  • comparison of the results of individual studies in the FAMRC to external studies 
  • comparison of reported to expected results. 

For example, we identified genetic variants that are known to cause cancer and checked that the same variants had the expected relationship in the FAMRC. In the figure below, every data point represents a single genetic variant that is known to increase cancer risk. The horizontal or X axis shows the known relationship in the GWAS catalog (this is a database of  known genetic associations with 1000s of human characteristics in 1000s of genetic studies) and the vertical or Y axis shows the relationship in one of the studies in the FAMRC. Each axis shows the Z score, which is basically a standardised measure of how each genetic variant affects cancer risk (positive values mean that the variant increases risk of cancer and negative values indicate they decrease risk). As you can see, in the FAMRC study on the vertical Y axis, almost all the variants have negative Z values (indicating they reduce cancer risk), when in fact they are known to increase risk (the true relationship is represented by Z scores in the GWAS catalog). This discrepancy was caused by a metadata error, where the effect allele column was incorrectly labelled. We also found that the “frequency of the effect allele” was wrong. How common the allele is in the population was opposite to what we’d expect, based on comparison with other studies, confirming the presence of metadata errors. 

Various other types of errors were identified, including one study reporting that 100s of genetic variants had very strong effects on fatty acid levels when in fact they had no effect at all.  For example, in the figure below, the many red data points refer to genetic variants in the FAMRC that had a very large effect on fatty acids but were not reported in the GWAS catalog, suggesting a potential problem with the genetic results.

We also compared the reported results (how the genetic variants affected fatty acids in the FAMRC) to predicted results (how we would expect the genetic variants to affect fatty acids). In the figure below we see a “fanning-out” pattern, when what we should see is a strong linear relationship (i.e. the data points lying on a single straight line). This relationship can be summarised with the “slope” metric. We should see a slope of 1 (this means if the reported result increases by 1 the predicted result will also increase by 1), which is not the case. We confirmed with the data provider that low quality genetic variants had not been excluded from their study. Once the low quality variants had been excluded, the discrepancies disappeared.

Avoiding metadata errors: recommendations for researchers

When conducting Mendelian randomization analyses using results from genetic studies, researchers can avoid metadata and other errors by: 

  1. Requesting results for genetic variants that are known to affect their disease of interest. Researchers should check that these variants have the expected effect in their dataset. 
  2. Comparing the frequency of genetic variants to expected frequencies in a reference dataset. We created a special reference dataset that can be used for this purpose (accessible via the CheckSumStats R package). 
  3. Not assuming that results have had low quality variants excluded, but instead seeking confirmation of this with data providers. Our quality control tools also provide a way to check this. 

Further attention is needed to address the growing diversity of GWAS

One issue we only partly addressed was the “two-sample assumption”: that the studies being compared come from the same population. In our own analyses, we found that the frequency of genetic variants was very similar across European-origin studies, indicating satisfaction of the assumption. On the other hand, our tools were not really optimised for this purpose. The need to assess the “same population” assumption is becoming more urgent with the growing diversity of genetic studies. 

In conclusion, meta and summary data errors are an under-appreciated source of bias in MR research, especially in complex study designs. We developed an R package and set of tools that can be used to flag meta and summary data errors in the results of genetic studies, which in turn can be used to enhance the integrity of Mendelian randomization analyses. Our tools and methods are available to other researchers via the CheckSumStats R package.  

Further reading

Design and quality control of large-scale two-sample Mendelian randomisation studies

https://www.medrxiv.org/content/10.1101/2021.07.30.21260578v1 

CheckSumStats package

https://github.com/MRCIEU/CheckSumStats 

Contact the author

philip.haycock@bristol.ac.uk 

Does testosterone drive success in men? Not much, our research suggests

Posted on by lindsey.pike
Not quite what the science says.
UfaBizPhoto/Shutterstock

Amanda Hughes, University of Bristol; Neil Davies, University of Bristol, and Sean Harrison, University of Bristol

There’s a widespread belief that your testosterone can affect where you end up in life. At least for men, there is some evidence for this claim: several studies have linked higher testosterone to socioeconomic success. But a link is different to a cause and using DNA, our new research suggests it may be much less important for life chances than previously claimed.

In previous studies, male executives with higher testosterone have been found to have more subordinates, and financial traders with higher testosterone found to generate greater daily profits. Testosterone has been found to be higher among more highly educated men, and among self-employed men, suggesting a link with entrepreneurship. Much less is known about these relationships in women, but one study suggested that for women, disadvantaged socioeconomic position in childhood was linked to higher testosterone later in life.

The beneficial influence of testosterone is thought to work by affecting behaviour: experiments suggest that testosterone can make a person more aggressive and more risk tolerant, and these traits can be rewarded in the labour market, for instance in wage negotiations. But none of these studies show definitively that testosterone influences these outcomes because there are other plausible explanations.

Rather than testosterone influencing a person’s socioeconomic position, it could be that having a more advantaged socioeconomic position raises your testosterone. In both cases, we would see a link between testosterone and social factors such as income, education and social class.

There are plausible mechanisms for this too. First, we know that socioeconomic disadvantage is stressful, and chronic stress can lower testosterone. Second, how a person perceives their status relative to others in society might influence their testosterone: studies of sports matches, usually between men, have often found that testosterone rises in the winner compared to the loser.

Older man holds head in front of a computer.
Chronic stress can lower your testosterone.
Shutterstock

It’s also possible that some third factor is responsible for the associations seen in previous studies. For instance, higher testosterone in men is linked to good health – and good health may also help people succeed in their careers. A link in men between testosterone and socioeconomic position could therefore simply reflect an impact of health on both. (For women, higher testosterone is linked to worse health, so we would expect an association of higher testosterone and lower socioeconomic position.)

Look at it this way

It is very difficult to pick apart these processes and study just the effects of testosterone on other things. With this goal in mind, we applied a causal inference approach called “Mendelian randomisation”. This uses genetic information relevant to a single factor (here, testosterone) to isolate just the effect of that factor on one or more outcomes of interest (here, socioeconomic outcomes such as income and educational qualifications).

DNA visualisation with coloured bars
DNA can tell us a lot about our relationship with testosterone.
Zita/Shutterstock

A person’s circulating testosterone can be affected by environmental factors. Some, like the time of day, are straightforward to correct for. Others, like somebody’s health, are not. Crucially, socioeconomic circumstances could influence circulating testosterone. For this reason, even if we see an association between circulating testosterone and socioeconomic position, we cannot determine what is causing what.

This is why genetic information is powerful: your DNA is determined before birth and generally does not change during your lifetime (there are rare exceptions, such as changes which occur with cancer). Therefore, if we observe an association of socioeconomic position with genetic variants linked to testosterone, it strongly suggests that testosterone is causing the differences in socioeconomic outcomes. This is because influence on the variants of other factors is much less likely.

In more than 300,000 adult participants of the UK Biobank, we identified genetic variants linked to higher testosterone levels, separately for men and women. We then explored how these variants were related to socioeconomic outcomes, including income, educational qualifications, employment status, and area-level deprivation, as well as self-reported risk-taking and overall health.

Similar to previous studies, we found that men with higher testosterone had higher household income, lived in less deprived areas, and were more likely to have a university degree and a skilled job. In women, higher testosterone was linked to lower socioeconomic position, including lower household income, living in a more deprived area, and lower chance of having a university degree. Consistent with previous evidence, higher testosterone was associated with better health for men and poorer health for women, and more risk-taking for men.

However, there was little evidence that genetic variation related to testosterone affected socioeconomic position at all. In both men and women we detected no effects of genetic variants related to testosterone on any aspect of socioeconomic position, or health, or risk-taking.

Because we identified fewer testosterone-linked genetic variants in women, our estimates for women were less precise than for men. Consequently, we could not rule out relatively small effects of testosterone on socioeconomic position for women. Future studies could examine associations in women using larger, female-specific samples.

But for men, our genetic results clearly suggest that previous studies may have been biased by the influence of additional factors, potentially including the impact of socioeconomic position on testosterone. And our results indicate that – despite the social mythology surrounding testosterone – it may be much less important for success and life chances than earlier studies have suggested.The Conversation

Amanda Hughes, Senior Research Associate in Epidemiology, University of Bristol; Neil Davies, Senior Research Fellow, University of Bristol, and Sean Harrison, Systematic Reviewer, University of Bristol

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Study collecting the views of young people, parents of children with long COVID, and doctors, finds that long COVID in children is poorly understood by doctors

Posted on by lindsey.pike

Dr Katharine Looker

‘Enhancing the utilization of COVID-19 testing in schools’, is a study which will look at the characteristics of long COVID and COVID-19 infection in children. ‘Long COVID’ is commonly used to describe signs and symptoms that continue or develop after acute COVID‑19. The study is being funded as a result of a rapid funding call by Health Data Research UK (HDR UK), the Office for National Statistics (ONS) and UK Research and Innovation (UKRI). The study forms part of the larger Data and Connectivity National Core Study, which is led by HDR UK in partnership with ONS.

The COVID-19 testing in schools study is related to the CoMMinS (COVID-19 Mapping and Mitigation in Schools) study being undertaken by the University of Bristol in partnership with Bristol City Council, Public Health England [PHE] and Bristol schools. CoMMinS aims to give us an understanding of COVID-19 infection dynamics centred around school pupils and staff and onward transmission to family contacts, using regular testing. Our study will jointly analyse data from CoMMinS, along with information from Electronic Patient Records, and data from the COVID-19 Schools Infection Survey (SIS; jointly led by the London School of Hygiene & Tropical Medicine [LSHTM], PHE, and ONS). The SIS is a study similar to CoMMinS but carried out nationally.

To help inform research questions and methods for the study, members from the University of Bristol study team gathered views about long COVID in children between 9 March and 30 April 2021 from:

  • seven young people from the NIHR Bristol Biomedical Research Centre Young People’s Advisory Group (YPAG)
  • five families whose children have long COVID or suspected long COVID, recruited through two online UK campaign groups for long COVID, and
  • a survey completed by four GPs and one paediatrician, and an online meeting with two paediatricians.

It is important to note that the opinions gathered were based on small samples which may not be representative.

Through the meeting and survey with the doctors, the study team found that clinical understanding of long COVID in children is currently very limited.

The doctors said that it may be hard to distinguish between long COVID and other conditions with similar symptoms. Many of the symptoms of long COVID, like fatigue and feeling sick, aren’t very specific, and are common to many different conditions. Long COVID in children currently lacks a clinical definition, making diagnosis difficult. It isn’t yet properly understood whether long COVID is a new condition in itself, or a group of conditions like post viral fatigue, which is already recognised.

Young people, and families of children with long COVID or suspected long COVID, who were also asked for their opinion, said that feeling sick or stomach pain, extreme tiredness, and headaches were the symptoms they would rank as most ‘harmful’. For young people, this was based on them imagining having the symptoms. For the families, this was based on their first-hand experience.

The families also said that the symptoms their children were experiencing were numerous, often very severe, and more wide-ranging than those currently listed on the NHS website for long COVID. It is not yet clear what is causing the unusual symptoms.

The families said that they had struggled to get a diagnosis and treatment for their children. They also said that long COVID symptoms were having a significant impact on their children’s day-to-day lives both physically and psychologically, and that some of the children had missed school because of the symptoms. Some of the families also found fevers difficult to manage because their children had to miss school to self-isolate every time they had a fever. They wanted to know why the set of symptoms were being experienced, and why their children in particular had developed them.

It is not known how many children have or will develop long COVID. So far, studies which have tried to measure the rate of long COVID in children suggest it is rare. However, quantifying the number of cases is made difficult by a lack of clinical understanding of long COVID including the lack of an agreed clinical definition. The opinions collected suggest that relying on clinical diagnoses alone will under-estimate cases. On the other hand, there needs to be a cautious approach to estimating the number of cases based on non-specific symptoms, as other conditions which cause similar symptoms may be counted as well.

Caroline Relton, Professor of Epigenetic Epidemiology and Director of the Bristol Population Health Science Institute at the University of Bristol, joint lead for CoMMinS and one of the lead authors of the report, said: “The opinions we gathered further highlight that it is difficult to count the number of children with long COVID on the basis of diagnoses alone while long COVID in children remains poorly defined.

“There are added complications of studying long COVID in children, when it is sometimes difficult to disentangle what might be the result of experiencing infection from what might result from the wider impact of experiencing the pandemic. Isolation, school closures, disrupted education and other influences on family life could all have health consequences. Defining the extent of the problem in children and the root causes will be essential to helping provide the right treatment and to aid the recovery of young people who are suffering.”

The findings highlight that examining GP and hospital visits, and school attendance, might currently be a more useful and feasible way of assessing how COVID-19 has affected children, rather than relying only on diagnoses of long COVID. However, the study researchers also need to be aware how often healthcare is accessed according to need, and absence from school due to self-isolation, which will affect what is being measured.

Feeling sick or stomach pain, extreme tiredness, and headaches will be important symptoms to consider in the study.

Read the full report

Find the full report on the CoMMinS study news page.

 

How can your gut microbiome affect risk of cancer?

Posted on by lindsey.pike

Dr Kaitlin H. Wade1,2,3

1 Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, BS8 2BN

2 Medical Research Council (MRC-IEU), University of Bristol, Bristol, BS8 2BN

3 Cancer Research UK (CRUK) Integrative Cancer Epidemiology Programme (ICEP), University of Bristol, Bristol, BS8 2BN

The causes of cancer are often preventable

Cancer, a disease that has a profound impact on the lives of individuals all over the world, also has an ever-increasing burden. And yet, evidence indicates that over 40% of all cancers are likely explained by preventable causes. One of the main challenges is identifying so-called ‘modifiable risk factors’ for cancer – aspects of our environment that we can change to reduce the incidence of disease.

Photo by Chloe Russell for ‘Up Your A-Z’, an encyclopaedia of gut bacteria

The gut microbiome could influence cancer risk

The gut microbiome is a system of microorganisms that helps us digest food, produce essential molecules and protects us against harmful infections. There is growing evidence supporting the relationship between the human gut microbiome and risk of cancer, including lung, breast, bowel and prostate cancers. For example, experiments have shown that changing the gut microbiome (e.g., by using pre- or pro-biotics) may reduce the risk of developing colorectal cancer. Research also suggests that people with colorectal cancer have lower microbiota diversity and different types of bacteria within their gut compared to those without a diagnosis.

As the gut microbiome can have a substantial impact on their host’s metabolism and immune response, there are many biological mechanisms by which the gut microbiome could influence cancer development and progression. However, we don’t yet know how the gut microbiome can do this.

Human studies in this context have used small samples of individuals and measure both the microbiome and disease at the same time. These factors can make it difficult to tease apart correlation from causation – i.e., does variation in the gut microbiome change someone’s risk of cancer or is it the existence of cancer that leads to variation in the gut microbiome? This is an important question because the main aim of such research is to understand the causes of cancer and how we can prevent the disease. We want to fully understand whether altering the gut microbiome can reduce the burden of cancer at a population level or whether it is simply a marker of cancer itself.

I’m inviting feedback on your knowledge and understanding of the gut microbiome and cancer – please take this 5-minute survey (click here for survey) to contribute your thoughts.

People are interested in their gut microbiome

Even though we don’t yet know much about the causal relevance of the gut microbiome, there is still a growing market for commercial initiatives targeting the microbiome as a consumer-driven intervention. This usually involves companies obtaining a small number of faecal samples from consumers and prescribing “personalised” nutritional information for a “healthier microbiome”. However, these initiatives are very controversial given uncertainty in the likely relationships between the gut microbiome, nutrition and various diseases. What these activities do highlight is the demand for such information at a population level. This shows there is an opportunity to improve understanding of the causal role played by the gut microbiome in human health and disease.

Photo by Chloe Russell for ‘Up Your A-Z’, an encyclopaedia of gut bacteria

Microbiome and variation in our genes

Using information about our genetics can help us find out whether the gut microbiome changes the risk of cancer, or whether cancer changes the gut microbiome. Genetic variation cannot be influenced by the gut microbiome nor disease. Therefore, if people who are genetically predisposed to having a higher abundance of certain bacteria within their gut also have a lower risk of, say, prostate cancer, this would strongly suggest a causal role of those bacteria in prostate cancer development. This approach of using human genetic information to discern correlation from causation is called Mendelian randomization.

Studies relating human genetic variation with the gut microbiome have proliferated in recent years. They have provided evidence for genetic contributions to features of the gut microbiome including the abundance or likelihood of presence (vs. absence) of specific bacteria. This knowledge has given the opportunity to apply Mendelian randomization to better understand the causal impact of gut microbiome variation in health outcomes, including cancer. There are, however, many important caveats and complications to this work. Specifically, there is a (currently unmet) requirement for careful examination of how human genetic variation influences the gut microbiome and interpretation of the causal estimates derived from using Mendelian randomization within this field.

This is exactly what I will be looking at in my new research funded by Cancer Research UK. For more details on the nuances of this work, please see my research feature for Cancer Research UK and paper discussing these complexities.

What’s next for this research?

This research has already shown promise in the application of Mendelian randomization to improve our ability to discern correlation from causation between the gut microbiome and cancer. It has importantly highlighted the need for inter-disciplinary collaboration between population health, genetic and basic sciences. Thus, with the support from my team of experts in microbiology, basic sciences and population health sciences, this Fellowship will set the scene for the integration of human genetics and causal inference methods in population health sciences with microbiome research. This will help us understand the causal role played by the gut microbiome in cancer. Such work acts as a new and important step towards evaluating and prioritising potential treatments or protective factors for cancer prevention.

Acknowledgements

The research conducted as part of this Cancer Research UK Population Research Postdoctoral Fellowship will be supported by the following collaborators: Nicholas Timpson, Caroline Relton, Jeroen Raes, Trevor Lawley, Lindsay Hall and Marc Gunter, and my growing team of interdisciplinary PhD students and postdoctoral researchers. I’d also like to thank the following individuals for comments on this feature: Tom Battram, Laura Corbin, David Hughes, Nicholas Timpson, Lindsey Pike and Philippa Gardom. Additional thanks go to Chloe Russell, a brilliant photographer with whom I collaborated to create “Up Your A-Z” as part of Creative Reactions 2019, who provided the photos for this webpage.

About the author

Dr. Wade’s academic career has focused on the integration of human genetics with population health sciences to improve causality within epidemiological studies. Focusing on relationships across the life-course, her work uses comprehensive longitudinal cohorts, randomized controlled trials and causal inference methods (particularly, Mendelian randomization and recall-by-genotype designs). Kaitlin’s research has focused on understanding the relationships between adiposity and dietary behaviours as risk factors for cardiometabolic diseases and mortality. Having been awarded funding from the Elizabeth Blackwell Institute and Cancer Research UK, Kaitlin’s work uses these methods to understand the causal role played by the human gut microbiome on various health outcomes, such as obesity and cancer. Since pursuing a career in this field, Kaitlin has already led and been key in several fundamental studies that with path the way to resolve – or at least quantify – complex relationships between genetic variation, the gut microbiome and human health. In addition to her research, Kaitlin is actively involved in organising and administering teaching and public engagement activities as well as having many mentorship and supervisory roles within and external to the University of Bristol.

Key publications:

Hughes, D.A., Bacigalupe, R., Wang, J. et al. Genome-wide associations of human gut microbiome variation and implications for causal inference analyses. Nat Microbiol 5, 1079–1087 (2020). https://doi.org/10.1038/s41564-020-0743-8.

Kurilshikov, A., Medina-Gomez, C., Bacigalupe, R. et al. Large-scale association analyses identify host factors influencing human gut microbiome composition. Nat Genet 53, 156–165 (2021). https://doi.org/10.1038/s41588-020-00763-1.

Wade KH and Hall LJ. Improving causality in microbiome research: can human genetic epidemiology help? [version 3; peer review: 2 approved]. Wellcome Open Res 4, 199 (2020). https://doi.org/10.12688/wellcomeopenres.15628.3.