Table 2.
NLP-ML models trained on microarray sample descriptions can accurately infer annotations for samples from five different genomics exp types
RNA-seq | ChIP-seq | Methylation array | CGHA | Microarray | Total | |
---|---|---|---|---|---|---|
Adipose tissue | 7 | 8 | 10 | 3 | 10 | 38 |
Brain | 10 | 10 | 10 | 9 | 10 | 49 |
Colon | 6 | 10 | 10 | 5 | 9 | 40 |
Neural tube | 10 | 10 | 10 | 7 | 9 | 46 |
Muscle tissue | 9 | 0 | 10 | 0 | 10 | 29 |
Each row corresponds to one of five top-performing NLP-ML tissue models. The last column shows the total out of 50 that each of these models annotated correctly. Columns 2–6 show the number of samples (out of 10) from each experiment type that each tissue model annotated correctly.
RNA-seq RNA-seq of coding RNA, Methylation array methylation profiling by array, CGHA comparative genomic hybridization by array, Microarray transcription profiling by array.