|
@SusanneTilk | |||||
|
Really excited to share the first chapter of my PhD on how genome-wide linkage in cancer reduces the efficacy of selection via Hill-Robertson interference (HRI), with mentors @cd_mcfarland, @PetrovADmitri and @cncurtis. bit.ly/2lTDQnT (1/10)
|
||||||
|
||||||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
Here, we try to resolve why signals of negative selection are largely absent in cancer, yet pre-dominant in the human germ-line. (2/10)
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
We hypothesized that since tumors evolve asexually, under the constraints of genome-wide linkage, mutations are unable to be removed by negative selection (or favored by positive selection). (3/10)
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
Since linkage effects get stronger as mutation rates increase, tumors with elevated mutational burdens should be particularly inefficient at removing deleterious passengers due to genetic hitchhiking and Muller’s Ratchet. (4/10) pic.twitter.com/vVW4z6C6q1
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
To test this idea, we calculated dN/dS in tumors stratified by their mutational burden and observe that negative and positive selection attenuates as more mutations accumulate in tumors. (5/10) pic.twitter.com/WF1WIJ02mV
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
We see the same trend in somatic CNAs. Using a dN/dS-style statistic called dE/dI (fractional overlap of CNAs in [E]xonic versus [I]ntergenic/[I]ntronic regions) – we similarly see that selection on CNAs attenuates as the mutational burden increases. (6/10) pic.twitter.com/LUg6hlRJFk
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
Importantly, we see that attenuated selection on CNAs and SNVs is generic to tumor evolution. This pattern persists across broad and specific tumor sub-type categories. (7/10) pic.twitter.com/bKJ8vSbQUV
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
Finally, we demonstrate that a simple evolutionary model incorporating HRI can explain these observed patterns of selection and allowed us to estimate the mean fitness effects of passengers (~1%) and drivers (~18%) – much larger than previously estimated. (8/10)
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
Thus, this allowed us to estimate that in elevated mutational burden tumors (>95% of cancers), deleterious passengers accumulate and confer an individually-weak, but collectively-substantial fitness cost of ~40% that impacts tumor progression. (9/10)
|
||
|
|
||
|
Susanne Tilk
@SusanneTilk
|
16. ruj |
|
How successful tumors overcome this deleterious load is an exciting, open question. Similar to constraints in germ-line evolution, we think that preventing protein mis-folding is an important first step. We’d love to hear what you think! (10/10)
|
||
|
|
||
|
Chris D McFarland
@cd_mcfarland
|
16. ruj |
|
I also want to thank my graduate advisor @leonidmirny, who first got me interested in deleterious passengers and their role in cancer.
|
||
|
|
||