Wiki Page

Visit the wiki page describing all MLRepo learning tasks.

Available Tasks

Download a single file containing all available tasks

Bacteremia

bacteremia vs no bacteremia

Diet

Antibiotics

Age

IBD

Gender

Vaginal

Geography

Body Habitat

Cancer

healthy vs tumor biopsy, paired

Obesity

lean vs obese, mz/dz/mom

Diabetes

Cirrhosis

healthy vs cirrhosis

Interested in adding a new dataset or task? See our instructions

If you use this package, make sure to cite at

@article{vangay2019microbiome,
  title={Microbiome Learning Repo (ML Repo): A public repository of microbiome regression and classification tasks},
  author={Vangay, Pajau and Hillmann, Benjamin M and Knights, Dan},
  journal={Gigascience},
  volume={8},
  number={5},
  pages={giz042},
  year={2019},
  publisher={Oxford University Press}
}

mlrepo's People

Contributors

Stargazers

Watchers

mlrepo's Issues

details of OTU names?

Hi developer,

I am interested in the dataset in this repo. However, I can not find the corresponding taxonomy of each OTU.
For example, the bacteremia dataset has 1852 OTUs in otutable.txt, but with 228 taxonomy taxatable.txt, and there is no key column between the two tables.
May I know where can I have these data?

Thanks in advance!
-Ivy

OTU representative sequences for the Ravel task

Hi,

Thank you for this excellent resource!

Are the representative sequences for each OTU available for the Ravel set of tasks? The processed sequences that are provided in the fastq file do not seem to correspond to any of the OTUs as they have names like

>SRR063057_0 SRR063057.1 FU604JZ01CTHV2 length=228

which correspond to the sample IDs (SRR063057 is a sample ID).

Thank you in advance!

How to use the result of collapase-by-correlation?

Hello,
I am interested in using the cluster.by.correlation and collapse.by.correlation functions in MLRepo/example/lib/collapse-features.r to decrease the number of features (variables) to be fed into a PCA or cluster analysis. I was able to run the functions, and it returned the group assignment (variable 1 belongs to group 1, variable 2 belongs to group 2, ... etc.)

How can I use the group assignment in further analysis? Would it return a new dataset with collapsed variables (i.e. highly correlated variables are thinned into one)? Or do I pick one variable myself to be fed into downstream analysis, like I pick variable 1 out of variables 1, 5, 9 in group 2, for example?

I hope it makes sense? Looking forward to hearing from you.
Rie

Recommend Projects

knights-lab / mlrepo Goto Github PK

mlrepo's Introduction