Thursday, December 31, 2020

Forming Voltron: bringing new and disparate skills together to study cancer evolution -- 4 new post-docs for Theory Division

4 new post-docs have joined us in 2020 whose skills have a perfect amount of overlap to allow collaboration, but disparate enough so that we will constantly be learning from one another. Look out cancer and pathogens, we're coming for you.


I have been extremely lucky in the first few years of my time building a laboratory to have had three extremely kind, talented and passionate post-docs choose to work with us. Starting with Nara Yoon, who completed her PhD in the math department at CWRU and is now an Assistant Professor at Adelphi University; and then Inom Mirzaev, who was a joint post-doc between Theory Division and the MBI at Ohio State University, who now works as a machine learning engineer; and then Emily Dolson, who came from (and went back to) Michigan State University, where she is now an Assistant Professor of computer science.

These three young scientists (along with some talented students who deserve their own post, but this one is about post-docs) really kick-started the work here in Theory Division, and helped spread the word about our work. When I finally got some real funding, and was able to hire more than one post-doc at a time, I started looking to hire two, as I knew Dr. Dolson was leaving us for the greener pastures of a tenure track job. This funny thing seems to happen to me... once I realize a post-doc is leaving I start to get really nervous. I think: who in their right mind would come work with me? How can I possibly ever replace (post-doc X) who is leaving? 

These two questions are both ill-posed, but for different reasons:

1 - you can't replace (post-doc X), they are wonderful and unique in their own way, and 'replacing them' would stifle the growth of the lab anyways -- one needs new ideas, new skills, new faces...

2 - 'who would ever work with me??' is an impostor syndrome laden question -- but impostor syndrome is real, I ask myself this question every time (and still don't know the answer!)

So, I was shocked when Kyle Card, finishing his Phd in the Lenski lab, responded positively to my twitter DM asking if he was thinking about post-docs yet (as I had just read his excellent new paper: Historical contingency in the evolution of antibiotic resistance after decades of relaxed selection). 


Kyle came to visit right before COVID, and we hit it off immediately, and before his visit ended, he took me up on an offer of a post-doc.

At the same time, I was chatting with Luka Opasic, who was completing his PhD with Arne Traulsen at the Max Planck Institute for Evolutionary Biology. Luka had spent a summer with us during the middle of his PhD (some of his work can be read here: How many samples are needed to infer truly clonal mutations from heterogenous tumours?), and was a great fit. A varied background in medicine and maths, and interest in evolution, and all the right kinds of nerd :).

The summer before all this happened, I was in Cambridge lecturing to the PhD students in a unique program called CONTRA: Computational ONcology TRaining Alliance Innovative Training Network.

After my lecture, I was approached after my talk (which you can watch here) by Steph Owen, a talented PhD student working in cancer who was eager to get back to their roots in Physics. As a physicist myself was now doing cancer research, we hit if off immediately.  The timing was such however, that Steph's PhD hadn't finished, and so we postponed the discussion of future work for a few months... which put our renewed conversation right at the beginning of COVID lock downs. Some of Steph's work can be read here: Combining measures of immune infiltration shows additive effect on survival prediction in high-grade serous ovarian carcinoma.

At this point there were three people interested when I imagined I wouldn't find any, and now I was nervous about how to make this happen (strictly from a funding perspective). Then, the universe spoke.  Nikhil Krishnan, a fourth year CWRU medical student in our lab (to whom I had promised PhD funding), got a Gates Fellowship to study Physics at Cambridge with Diana Fusco -- freeing up the funds to hire another person -- we'll call them the Nikhil Krishnan Post-doctoral fellow :)

The final piece of the puzzle I didn't know I was building was an unlooked for applicant. A friend and colleague, Kevin Wood from UM Biophysics visited us (actually the same weekend Kyle Card interviewed) to give a Biophysics colloquium, hosted by me and Mike Hinczewski (a friend and collaborator in CWRU Physics). A grad student in his lab (essentially Black Mirror version of our own lab), Jeff Maltas, was just finishing his PhD. It turns out that we had reviewed half of his thesis papers (anonymously) so I was quite familiar with his (excellent!) work (I take some credit for Figure 6 :) in: Pervasive and diverse collateral sensitivity profiles inform optimal strategies to limit antibiotic resistance). I didn't think Jeff would want to work with us, but invited him for a visit anyways.  and then, there were 4...

I couldn't have possibly decided between the 4, so I decided to go with them all, and sort the funding out later. Turns out, they did it for me -- between a computational genomics T32 and a competitive NIH diversity supplement, the universe (and my talented new young scientists) provided, and I am SO EXCITED to see what science we do together. There is just enough overlap between us all that we can speak freely and in our own scientific mother tongues, and enough separation that each is synergistic with the others. I feel incredibly honored to be able to host these young scientists and can't wait to see what science we're able to in Theory Division in 2021.



Wednesday, December 16, 2020

The effect of genetic background on the evolution of antibiotic resistance.


Our second video is a long form research talk from one of our awesome new post-docs, Kyle Card. Kyle came to us from Richard Lenski's lab at Michigan State University, and is working on disentangling population size, mutation rate and genetic background in the evolution of antibiotic resistance.  Check out his talk here:

https://www.youtube.com/channel/UCr_3HE-tUGu9mlSHQGZSYQA

Short Abstract:

Antibiotic resistance is a growing public-health concern. Efforts to control the emergence and spread of resistance would benefit from an improved ability to forecast when and how it will evolve. To predict the evolution of resistance with accuracy, we must understand and integrate information about many factors, including a bacterium’s evolutionary history. In this presentation, I discuss the effects of genetic background on the evolution of phenotypic resistance, its genetic basis, and its fitness costs.

First, I ask how readily bacteria can overcome prior losses of intrinsic resistance through subsequent evolution when challenged with antibiotics. Second, I consider whether lineages founded from different genotypes take parallel or divergent mutational paths to achieve increased resistance. Third, I inquire whether fitness costs of resistance mutations are constant across different genetic backgrounds. In general, I focus attention on the interplay between repeatability and contingency in the evolutionary process. I demonstrate that genetic background can influence both the phenotypic and genotypic evolution of resistance and its associated fitness costs. I conclude this presentation by highlighting clinical and public-health implications of this work.

Wednesday, December 9, 2020

Introducing our new Theory Division YouTube channel!

It turns out managing a blog is really hard during the assistant professor years.  As in, I've posted on average once a year since I started this job...  

So -- in hopes of doing better, we're going to start something new!

Each week in Theory Division we have lab meeting (which in itself is worthy of a post -- turns out no one teaches you how to run these, and how to best do it changes a LOT depending on the size of your lab), and we also have a one hour slot separately for either journal club, or for a long form research talk. Given the pandemic, these have been done via Zoom, and so have been recorded.  So, what we are going to do is, for each long form research talk we will post the recording to our new youtube channel, along with a short textual abstract here on the blog with a link to the talk.



I'll start things off!

Just yesterday, my colleagues and I had a paper come out in the Journal of Thoracic Oncology: 

Personalizing Radiotherapy Prescription Dose Using Genomic Markers of Radiosensitivity and Normal Tissue Toxicity in Non-Small Cell Lung Cancer

This paper was the subject of my Peter Canham lecture in biophysics at Western University this year -- and we have ported this talk to be our first in the channel -- head over to the channel to watch/subscribe!

Click here to watch this talk on our new channel.

In this talk I discuss some of the frustration I feel as a radiation oncologist that our field has not yet entered the era of personalized medicine -- that is, each cancer patient doesn't have their radiation therapy (DOSE!) tailored to their tumor's genomic profile. Certainly, each prescription is physically personalized (geometrically), but biologically personalized dosing is not yet standard of care. Starting with the beautiful opening (to use a chess analogy) starting with a breakthrough creation of a Radiation Sensitivity Index (RSI) which linked canonical ideas of SF2 in radiation biology to genomics, by my friend and mentor Javier Torres-Roca, he and I and our colleagues have slowly and methodically moved through the middle game, setting the stage for a relationship to dose, which we outlined in our article: A genome-based model for adjusting radiotherapy dose (GARD): a retrospective, cohort-based study

We then (I hope) end the middle game with this paper from yesterday (see above), where we use information about what dose is needed to optimize a given patient's tumor control to then make inferences about OVER and UNDER treatment -- and using a novel combined model of TCP and NTCP, determine how much we could make radiation therapy better right now.

So - head on over to our youtube channel, watch the video, let me know what you think!  And subscribe to see future talks in our lab!