DESeq2 removing columns from model matrix implication
Entering edit mode
Last seen 3.0 years ago

It is my understanding that when making comparisons both within and between groups with paired samples, but there are a differing number of samples for each condition, the model matrix must be adjusted as such: C: DESeq2 paired and multi factor comparison


I am confused as to what removing elements from the model matrix means. Does this mean that I am losing information from paired samples that don't have a counterpart in the other condition? Am I throwing away information?

I apologize if my question is phrased poorly--I'm pretty confused and would appreciate any clarification. Thank you!


deseq2 multifactorial design paired test • 561 views
Entering edit mode
Last seen 4 hours ago
United States

What's happening is that you are removing columns, corresponding to coefficients to estimate which have no samples for which to estimate those coefficients. Hence, the model matrix has a column of zeros. Ones in the column correspond to samples which will be given a 'beta' in the formula.

So it's definitely not throwing away information. If you tried to fit the model without removing those columns, the software would stop and say some version of: "I can't fit this model because there are no samples to estimate one or more coefficients."

Entering edit mode

Great, thank you so much for the clarification


Login before adding your answer.

Traffic: 166 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6