2
$\begingroup$

This is the data I have now: 30 simple sequence repeat (SSR) markers for 80 cultivars of cucumber. 10 of the 80 cultivars belong to one cultivar (let's say A).

My goal is to classify an unknown cultivar into A or "not A" using tge 30 SSR markers.

I'm think of considering it as a classification problem (A vs non-A) and use machine learning method to build a model using the SSR markers as features. But the problem is that A cultivars don't have enough number of samples.

Do you have any suggestions which statistical method(s) I can try to solve this problem? Thanks in advance.

$\endgroup$
2
$\begingroup$

Not to discount @Michael's answer (unsupervised learning is very handy for descriptive analysis of this kind of problem), but classification based on microsatellite data should not be too difficult. For papers with examples, you can see e.g. this and this. They seem to apply various algorithms, you might be able to learn something from those papers.

More generally, I bet that you could get pretty far with just (multinomial) logistic regression for your outcome variables.

Of course, this depends to a certain extent on how you code the data and what the input data look like. Microsatellite allele coding and evolutionary/error models are a field in themselves.

$\endgroup$
2
$\begingroup$

Use Fst (migration statistic) such as the Fstat program to generate a distance matrix and solve it via clustering, such as UPGMA or neighbor-joining. This analysis is particularly useful for diploids. I presume you have heterozygotes in your analysis. You can switch Fis (inbreeding) if you have excessive homozygosity.

The other way in is a haplotype network. When I get a minute I'll post the link to the programs. Machine learning isn't needed in this instance, although you could do a PCA style analysis, i.e. unsupervised learning. Unsupervised learning doesn't need a large sample size.


The other solution is a stepwise mutation phylogeny, but this is for microsatellites (one to four repeats). This is called Delta Mu squared and describes the method of repeat unit mutation. It is a distance method.

Again when I have a moment I will supply weblinks.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.