Halfpipe: a tool for analyzing metabolic labeling RNA-seq data to quantify RNA half-lives

We introduce Halfpipe, a tool for analyzing RNA-seq data from metabolic RNA labeling experiments. Its main features are the absolute quantification of 4sU-labeling-induced T>C conversions in the data, calculating the proportion of newly synthesized transcripts, and estimating (compartment-specific) RNA half-lives. Halfpipe excels at correcting critical biases caused by typically low labeling efficiency. We measure and compare the RNA metabolism in the G1 phase and during the mitosis of synchronized human cells. We find that RNA half-lives of constantly expressed RNAs are similar in mitosis and G1 phase, suggesting that RNA stability of those genes is constant throughout the cell cycle. Our estimates correlate well with literature values and with known RNA sequence features. Overall design: HeLa-S3 cells were cultured at 37°C in DMEM supplemented with 10% FBS Superior (Sigma, S0615) and 1X penicillin-streptomycin (Gibco, 15070063). A double thymidine block was used for cell cycle synchronisation at the G1/S boundary. Briefly, Hela-S3 cells were blocked with 2.5 mM thymidine for 18 hours. The cells were released from the block by two washes with pre-warmed medium, followed by incubation for 9 hours in fresh medium supplemented with 24 µM 2-deoxycytidine. The second G1/S block was again performed by adding 2.5 mM thymidine for 18 hours. Finally, cells were released by two washes with pre-warmed media and incubated for 6 hours (late G2) in DMEM containing 24 µM 2-deoxycytidine before labelling with 500 µM 4sU (Glentham Life Sciences, GN6085) for 0, 15, 45, 75, 120 and 180 minutes. At each time point, 3 million cells were fractionated to obtain nuclear and cytosolic fractions based on Mayer et al. (2015) with the following modifications: Nuclear pellets were washed with nuclear wash buffer (PBS with 1X cOmplete Protease Inhibitor (Roche, 11873580001)) and resuspended in 1400 µl TRIzol (Invitrogen, 15596026). Similarly, a time series in the G1 cell cycle phase (12 hours release time after double thymidine blocking) was measured at 0, 15, 45, 75, 145, 180, 240 and 300 minutes. Prior to phenol extraction of RNA, 1.9 µl of 1:100 diluted ERCC ExFold RNA Spike-In Mixes (Invitrogen, 4456739) per 1 million cells was added to each sample. The 4sU RNA was alkylated according to Herzog et al. (2017), and poly-A enriched SLAM-seq libraries were generated using the 3'-QuantSeq-REV protocol (Lexogen) according to the manufacturer's protocol. The final libraries were sequenced on an Illumina NovaSeq6000 sequencer at 150 nucleotide read lengths in paired-end mode.