Hello everyone,
I am looking for an open dataset to verify the results obtained by analyzing the total RNA-seq of patients in the TARGET-AML project (GEO search was not successful).
The dataset should include:
1) bone marrow RNA-seq of pediatric patients with non-relapsed AML;
2)bone marrow RNA-seq of pediatric patients with relapsed AML (primary tumor BEFORE relapse);
3)clinical data (relapse-free time, for example) - optional.
If you know where to find a dataset like this, I would be very thankful.
If you have a real sample I can help you to create a synthetic data to increase your sample size based on some machine learning methods.
Alexander S. Luchinin Hi, I'm very interested in your idea mentioned above. Why do you adapt machine learning to RNA-seq analysis? How does your machine learning method work in RNA-seq analysis (or simple descriptions step by step or conclusions are both OK)? What are the potential advantages of your machine learning methods?
Thanks a million.
Wei Zhou
By using the method, you can create a new dataset based on your real dataset. For example, it can be used to increase the size of a sample. It works only with numbers.
Alexander S. Luchinin Hi, thanks for your immediate reply. If I well understand your words, the methods you take advantage of will simulate the reads/counts of RNA-seq in virtual numbers. However, the questions also follow: 1. What are your trained datasets? Does the data fit any RNA-seq data? 2. The data/sample produced by your methods is convinced? Could you give more examples your methods applied?
Yours,
Wei Zhou
Wei Zhou Not exactly. By using the method, you can create new data based on your existing data. There is no previous trained model. It is a real-time algorithm. As I cannot see your data exactly, I am unable to tell you if this approach could help. Imagine, for example, a dataset of blood tests with n rows and k columns. By using this method, the next n rows can be created based on the previous ones. New data will be distributed similarly to primary data.
Alexander S. Luchinin Wow, that's cool! I think your method based on machine learning could apply to more scenario such as single sample GSEA (ssGSEA), CRISPR-screening, and other transcriptome data with only one samples. By the way, I would like to know whether your methods have been published on the journal or have some scripts on the github? If you could share the doi or github links, I would much appreciate it.
Wei Zhou
Wei Zhou Article Anonymiced Shareable Data: Using mice to Create and Analyze ...
- Badges
- Science topic