Fig. 1

Relationship between the number of reads and assembled transcript count across multiple cells and tissues. Transcript assembly and coding potential evaluation were done for each of the 671 short-read samples from multiple cells and tissues. The relationships between the number of reads and the transcript count for all transcripts (A), coding transcripts (B) and noncoding transcripts (C). For panels A- C and F–H, each point represents a sample while the line represents the loess fit with standard error band shown. Rho is the Spearman rank correlation coefficient. The distributions of the samples are shown on the axes. D. Number of reads versus coding to noncoding ratio. E. Increment in the number of transcripts per additional 10 million reads. The numbers of transcripts from 0 to 240 million reads with 10 million increments were predicted from loess fits in panels A-C. The number of transcripts due to additional 10 million read increment in read depth is shown on the vertical axis. Numbers of reads versus the number of transcripts for all GENCODE-annotated transcripts (F), GENCODE-annotated coding transcripts (G) and GENCODE-annotated noncoding transcripts (H). Assembled transcripts that were not found in GENCODE annotation were not included I. Increment in the number of transcripts inferred from loess prediction per additional 10 million reads for GENCODE-annotated transcripts. J. The difference between the observed and the predicted numbers of transcripts based on the loess prediction of all the 671 samples. Cells or tissues with at least 8 biological or technical replicates in the dataset are presented. K. The coding to noncoding ratio varied across cells and tissue types. Only cells or tissues with at least 8 biological or technical replicates are shown