search

const-ae / proDA

Protein Differential Abundance for Label-Free Mass Spectrometry https://const-ae.github.io/proDA/
17 stars 8 forks source link

quick question: data with no replication #15

Closed hukai916 closed 2 years ago

hukai916 commented 2 years ago

Hi Constantin,

Seems that proDA can not be used for data with no replication. Do you have any recommendations on analyzing such data?

Thanks!

--Kai

const-ae commented 2 years ago

I would recommend trying to get replicates.

Otherwise, you have little chance to estimate the variation per protein and thus assess if an observed difference is meaningful. If acquiring replicates is impossible, I think the best you can do are qualitative comparisons, but nothing quantitative.

hukai916 commented 2 years ago

Thanks! For explorative purposes, do you think fake duplication would make any sense when analyzing the data?

const-ae commented 2 years ago

I am not sure what you mean by fake duplication, but it sounds sketchy. For training purposes, I would rather recommend working with some published data.

hukai916 commented 2 years ago

Thanks! One of our preliminary set only have 1 replicate per sample, I tried to feed it into proDA, however, proDA won't accept data that are without replication. Therefore, I am thinking whether it would make any sense to make proDA work by simply "manually replicating" the data (copy "col1" into "col1-1" and "col1-2" as fake duplicates).

After this "manipulation", proDA seems to be able to take the data and produce some results. Again, this is only for explorative purpose. Do you not recommend it at all even for explorative analysis? Thanks!