Gustavo Valadares Barroso

Logo

email: gvbarroso (at) gmail.com

I am an population geneticist working to extract meaningful information from data using mathematics. A wet-lab / field biologist by training, I grew interested in statistical modelling after my masters. Nowadays, I approach biological questions with computational methods, using a combination of the C++ and R programming languages.

View My GitHub Profile

About

I was born in Jaragua do Sul, Santa Catarina, Brazil.
As a teenager, I played competitive Chess and in my early 20’s I was a professional Poker player. At the moment I am quite busy recovering from the last season of Bojack Horseman

I did my Bsc in Biological Sciences, Msc in Genetics and Evolution and PhD in Evolutionary Biology (CV). Some of my sarcasm towards academia and politics can be found on Twitter, but you should probably send me a private message if you want the really spicy stuff.

Academic Life

I want to understand how population-level processes interact with molecular-level processes to shape patterns of genetic and phenotypic variation. To this end, I develop probabilistic models based on population genetics theory and apply them to both simulated and empirical datasets from a variety of species. My long-term goal is to build more realistic models of nature.

I defend Bayesian reasoning as well as hypothesis-driven research. I also advocate for soft lottocracy in grant & job selections and support PeerCommunityIn, a platform for sustainable scientific publishing.

PhD

In the first chapter of my PhD thesis, I used linear modelling to investigate the signature of selection in gene expression variation among individual cells. Using single-cell transcriptomics from Mus musculus, we found that expression noise is reduced in genes that are central within molecular networks, consistent with selection acting to avoid noise propagation within the cell.

In the second and third chapters I developed the integrated Sequentially Markovian Coalescent (iSMC) – a population genetic model that uses Hidden Markov Models to jointly infer demography and variation in recombination and mutation rates along the genome. Using iSMC, we were able to show that the similarity of the recombination landscape among Hominin populations recapitulates their established evolutionary history, as well as to estimate that mutation rate variation is the main driver of the distribution of nucleotide diversity along the fruit-fly genome. A thorough review of Coalescent Hidden Markov Models was written by my PhD advisor, Julien Dutheil.

Also during my PhD, we wrote a chapter in this open access book: Statistical Population Genomics.

Postdoc-ing

UCLA During my postdoc with Kirk Lohmueller I switched my focus to the population genetics of natural selection. I first developed TIDES – a novel statistical framework for inferring parameters of natural selection by levereging transmission distortions in large datasets of family trios. Our method uses approximate Bayesian computation to jointly infer the dominance and selection coefficients of mutations. TIDES is a very flexible model that makes minimal assumptions about the evolutionary process that shaped the DNA sequence data, and we are currently working on improving it further, both conceptually (e.g., by modelling epistasis) and practically (by including new summary statistics using reasonable computational resources) – stay tuned!

I am now using SLiM simulations to explore the factors that may limit the accuracy of more traditional methods for inferring selection parameters (i.e., those based on the Site Frequency Spectrum), and thanks to Chris Kyriazis I became interested in less mainstream (yet quite intriguing!) questions in population genetics such as the effect of age structure and life-history strategies on the efficiency of selection.

UW-Madison As of May 2022 I will be a postdoc with Aaron Ragsdale. There I will face my scariest demons and finally find out if I can do “proper” population genetics theory!

Life outside academia

Truth be told, these days life outside academia essentially means Chess. What started out as a little pandemic project quickly gained momentum and I am even coming back to the “competitive” scene. My long-term goal is to get the title of FIDE master, but that should take a good number of years and whether I can keep a decent study routine of the game remains to be seen. In the meantime, you cand find me playing blitz and rapid on Lichess under the nickname QuebelezaFirula, a Brazilian Portuguese wordplay with the name of the French-Iranian genius Alireza Firouzja.

Other than that, I very much enjoy cooking, playing tennis, football and – of course! – hiking. And if chess would not be nerdy enough, you can add Risk (especially the Game of Thrones variation) and Magic:the Gathering’s EDH format to the list of games that I take rather seriously.
Next on the bucket list is learning how to play an instrument. At 16, I tried the bass, but even though I had long hair life got in the way at the time…

By the way, the draft of my memoir is already 10 pages long!