Question: breast cancer subtype
1
gravatar for kanwarjag
3.7 years ago by
kanwarjag1.0k
United States
kanwarjag1.0k wrote:

Is there a simple way of classifying breast cancer subtype. Clustering will classify samples in terms of expression of set of genes. Then how one decide which type is this. I did my home work a lot but could not find a very straight forward answer. Any pointer may be helpful- publication/ guidelines how to classify

latest • 1.9k views
ADD COMMENTlink modified 18 months ago by leizhengdeng0 • written 3.7 years ago by kanwarjag1.0k
1

You are not going to be able to classify cancer subtypes without taking into consideration the clinical/pathology data that accompanies the samples (unless you have developed subtype specific signatures a priori). Perou lab at UNC has been working on this topic for the past 15+ years. Take a look at the publications list here

ADD REPLYlink modified 14 months ago by RamRS24k • written 3.7 years ago by genomax74k
1

I'll have to respectectfully disagree here. It is possible to classify breast cancers solely bases on transcriptomics data without any additional clinical data, such as immunohistochemical data.

The lab you cite has released its own signature assay based on the PAM50 gene signature.

In fact, that is the future and the whole point of next-generation sequencing in the field of cancer. Sequence the tumor, and identify the tumor subtype, and the appropriate treatment, without any further tests.

ADD REPLYlink modified 14 months ago by RamRS24k • written 3.7 years ago by ablanchetcohen1.2k

Would you like to point me out that if XXX genes are over expressed then which subtype is that particular type. I simply never found that information. perhaps I may be missing something

ADD REPLYlink written 3.7 years ago by kanwarjag1.0k

If you're interested in the PAM50 gene signature, which is one of the best known gene signatures to classify breast cancer subtypes, you should consult the following thread. It's a gene signature, incidentally, so the genes within the signature are just as likely to be underexpressed as overexpressed, or even have no change in gene expression. Where To Download Pam50 Gene Set?

ADD REPLYlink written 3.7 years ago by ablanchetcohen1.2k

I was referring to general cancer sub-typing in first sentence.

PAM50 signature was developed by a large amount of work that used other evidence. That is going to hold true for any cancer. But I do agree with your point about the long term use of these signatures once they are developed.

ADD REPLYlink written 3.7 years ago by genomax74k

Based on gene expression profiling, there are three subtypes of breast cancer: Mesenchymal, Proliferative, and Metabolic.

Mesenchymal will response to Tamoxifen; Proliferative will response to taxol and trastuzumab; Metabolic will response to 5-FU.

ADD REPLYlink written 18 months ago by leizhengdeng0
1
gravatar for ablanchetcohen
3.7 years ago by
ablanchetcohen1.2k
Canada
ablanchetcohen1.2k wrote:

No, there is no simple way of classifying breast cancers by subtype.

There are a number of algorithms that have been developed, some of which have been implemented in R packages.

You could check out genefu and citbcmst.

It isn't exactly my research topic, but I used to work with the developers of genefu.

It's just not a simple question.

I've had entire courses devoted to the question given by computational biologists who have spent their entire careers working on breast cancer classification.

If you're inclined to do so, you can try developing your own method, by following the tutorial posted on this website.

ADD COMMENTlink modified 14 months ago by RamRS24k • written 3.7 years ago by ablanchetcohen1.2k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1316 users visited in the last hour