DIYBIO.to Updates!

Dec. 20, 2024

As our planet spins around the sun, and we are almost turning into a new year (Solstice is tmro!), it’s time to take stock of the past year, and all the good stuff that has happened at DIYBIO.to this year.  

Bioinformatics group news 

Heraclitis is stated something like “Noone ever steps in the same river twice”, meaning that change is the only constant. This is nowhere more true than in bioinformatics. We are interested in providing some basic training and discussion in this area, as it is important for early stage science students to get a grasp of, but also for the public to be able to access and understand data of their own, or public data. This information is freely available, and as a community group that is interested in encouraging public engagement in biology and computers, we have meetups to encourage this. But to get through this we need to define our objectives clearly. 

Some of us at DIYBIO are interested in making bioinformatics for accessible using: 1. gene expression in datasets that are publicly available and 2. metabolic pathway changes in non-health related data (plant for instance) as an example of how any genetic expression data can be used. 

It is important to note that the source of the original data in health bioinformatics can influence how we interpret results. For instance, background data on gene ontology and pathways can preferentially weight oncologic data, due to the vast amount of data in that area being added to KEGG, webgestalt, PANTHER, DAVID and other sites. Even if analysis is weighted and an appropriate reference dataset is used, there is just so much more historical data in the cancer area that cancer pathways will likely be enriched and found in non-cancer disease. This can confound the ability to find unique pathways for unique understudied diseases. It can also potentially weight cancer studies toward the most studied cancers. For this reason, a more historical-data agnostic approach may be needed, such as a transcriptional approach. For instance, looking at the potential binding partners in upregulated genes, would be agnostic to historical pathway data on cancer. For example, cytohubba and iRegulon in R and Cytoscape, DIRE and Opossum could be used, which is what we are presently doing!

In other bioinformatics approaches, we are looking at the absolute latest in advances in high resolution spatial genomics and trying to tease apart the potential approaches in these new techniques. The data available online is astounding! But using it in downstream methods is relatively new and challenging. Despite this, we have already performed some initial analysis of public datasets using pseudobulk methods after running Seurat and SpaCET, then DESeq2. Ask at our weekly meetups, diybio email or the hacklab open house, if you are interested in contributing to this project.

Nov. 27, 2024