Benchmarking results under a 70% dropout rate.

<p>(A) 2D visualization of the masked dataset (seed = 88) before and after imputation, illustrating the restoration of data structure. (B) Pearson correlation coefficient (PCC) between the imputed and ground truth expression. (C) Root mean square error (RMSE) of imputed values. (D) Proportion...

Full description

Saved in:

Bibliographic Details
Main Author:	Siyi Huang (8562174) (author)
Other Authors:	Linfeng Jiang (2416375) (author), Ming Yi (15051) (author), Yuan Zhu (148570) (author)
Published:	2025
Subjects:	Genetics Molecular Biology Plant Biology Biological Sciences not elsewhere classified Mathematical Sciences not elsewhere classified Information Systems not elsewhere classified xlink "> single uses bulk rna recovers expression values providing clear guidelines including cell clustering false measurements caused extensive practical evaluation essential downstream analyses differential expression detection cell rna sequencing cell &# 8217 three key innovations guided imputation engine distort biological signals aware normalization step d3impute demonstrates consistent accurately identify non true biological zeros seq data analysis biological zeros true transcriptome key hurdle guide imputation biological reference aware modeling aware discrimination seq data inflated data data recovery trajectory inference technical limitations specific characteristics significant improvements oriented solution optimal application network discriminator major challenge introduce d3impute handling zero genuinely absent generalizable framework fundamental issue computational methods comprehensive benchmarking cellular heterogeneity biologically informed also offers 12 state
Tags:	Add Tag No Tags, Be the first to tag this record!

Benchmarking results under a 70% dropout rate.

Similar Items