class: title-slide center middle background-image: url("img/hands.png") background-position: right background-size: contain background-color: white .pull-left[ # .red[Hypothesis testing course summary] ### .red[Rachel Warnock] ### .red[12.04.2024] ] --- class: middle center background-size: contain background-position: middle background-image: url("reprod-study.png") --- class: middle <style type="text/css"> .large { font-size: 130% } </style> # Reproducibility in cancer biology .large[Aim: replicate (and reproduce) 193 experiments from 53 high-impact papers.] .large[In the end, they were able to repeat 50 experiments from 23 papers.] .large[The problems: * **None** of the 193 experiments were described in sufficient detail in the original paper to repeat the experiment. * Key descriptive details necessary for statistics available for only 4 of 193 experiments. * Despite contacting the authors, they were missing data for 68% of experiments.] .footnote[[Errington et al. 2021](https://elifesciences.org/articles/67995)] --- class: middle # Reproducibility in cancer biology .large[Authors were extremely or very helpful for 41% of experiments, minimally helpful for 9% of experiments, and not at all helpful (or did not respond to us) for 32% of experiments.] .large[Once experimental work started, 67% of the peer-reviewed protocols required modifications to complete the research and just 41% of those modifications could be implemented.] .footnote[[Errington et al. 2021](https://elifesciences.org/articles/67995)] --- class: middle background-size: contain background-position: right background-image: url("barriers.png") .pull-left[ .large["*It is hard to assess whether reported findings are credible.*"] .pull-left[ .footnote[[Errington et al. 2021](https://elifesciences.org/articles/67995)] ] ] --- class: middle # What about paleobiology? --- class: middle background-size: 80% background-position: centre background-image: url("img/proj_friday.png") --- <br> <br> .large[ *"When [...] datasets were not available [...], we sent a personalized template email [...] to the lead or corresponding author(s) of PBDB publications asking for the dataset. If no response was received after 2 weeks, we contacted authors again with a follow-up email."* ] -- .large[ *"Within a few days (median = 1 day, mean = 5.5 days), 50% of the 167 responses provided either a file or a link to the requested data, 17% of responses indicated that the files had been lost, 9% of responses indicated only simple use of the PBDB that required no download, and 23% of responses indicated the publication did not use PBDB data."* ] -- .large[ *"We did not receive a response from authors for 25% of the PBDB publications (68/ 268 requests). In total, we were able to extract the needed information from 151 PBDB publications, accounting for* **38% of PBDB publications** *(total = 396) within the temporal scope our data collection phase"* ] --- class: middle # What will *you* do to ensure your research is reproducibile? --- class: middle background-size: contain background-position: left background-image: url("at-first.png") .pull-right[ # Reproducibility in R .large[R and R Markdown (or other tools for organising code and notes) can help: * keeping track of progress * sharing progress (during & after your project) * enable reproducibility ] ] --- class: middle # Research design .large[A **hypothesis testing framework** can help guide your project design *objectively*, including choice of data, sampling approach, and statistical tests.] -- .large[You do not need remember the details of statistical tests! You only need to know where to look for them. Take advantage of available tools!] --- class: middle center inverse .pull-left[ <center><img src="https://media0.giphy.com/media/v1.Y2lkPTc5MGI3NjExeHAyZTA3Z3h1bXRxeG0yZmF0bnJsaWYxbnU2MGNjYTZxN2s2bXlwYiZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/ekwEeLxb7G4DW44YaK/giphy.gif" alt="" height="350px" /></center> ] .pull-right[ <br> <br> ## When in doubt, ask your peers! ] --- class: middle centre inverse .center[ .pull-left[ <br> <br> # Thank you! ] ] .pull-right[ <center><img src="https://media.giphy.com/media/Uv1ocOCpjNRDc3vZse/giphy.gif" alt="" height="300px" /></center> ] --- class: middle center In paleobiology we delve,<br> To learn about the creatures that once swelled,<br> But when we add statistics to the mix,<br> Our students' frustration reaches its peak. They count and measure and analyze,<br> But the data just confounds their eyes,<br> They long for fossils and ancient bones,<br> Not p-values and confidence zones. But fear not, dear students, don't despair,<br> Statistics is a tool that's fair,<br> It helps us find the truth we seek,<br> And in the end, the answers speak. So buckle down and crunch those numbers,<br> And soon you'll find your knowledge lumbers,<br> For in the intersection of these two fields,<br> Lies the secrets that evolution yields. *A poem by ChatGTP*