ProTInSeq: transposon insertion tracking by ultra-deep DNA sequencing applied to identify small and large translated ORFs

ProTInSeq is a novel -omics technique designed to characterize proteomes by using DNA ultra-deep sequencing. The technique is based on transposons engineered to have a positive or negative protein selection marker expressed when the transposon is inserted in-frame into a protein-coding gene. In the genome-reduced bacterium Mycoplasma pneumoniae, ProTInSeq identifies 80% of known expressed proteins, as well as 5 new open reading frames (ORFs; >100 amino acids); and 153 novel small ORF-encoded proteins (SEPs; ≤100 aa) that represent up to 18% of this bacterium’s proteome. ProTInSeq can be used to detect translational noise, for protein quantification and to provide insight into functional protein aspects such as relative half-life, stability, and membrane topology. Herein, we describe a methodology that can be easily implemented in any living system and allows the deep understanding of proteomes and more importantly the identification of small proteins by DNA ultra-sequencing.

We include the following files:

- processed_inscalling.zip: output obtain after running FASTQINS transposon calling tool over the datasets found at https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-10380?key=5f54209d-ce59-490a-9bcd-7084e9c619ee. This include every genome position in M. pneumoniae, the number of times an insertion has been mapped to that position and the total read count value.

- separated_library_metrics.zip: insertion and read count processed from processed_inscalling files associated to every ORF  and intergenic region in M. penumoniae. Columns include frame measured (0 - whole gene, 1 - in-frame, 2 and 3 for following positions) and metric. Metrics account for number insertions in-frame (I), read count (R), linear density from non-coding regions used in the Poisson evaluation (rNC), probability measured (sfNC) and a binary for prediction (pred; 0 - no significant, 1 - significant).

- allmetrics.xlsx: merged table with the combination of results from separated_library_metrics.zip tab-delimited files.

- selective_metrics_allannotations.xlsx:  this table includes all the available information about the 30,112 sequences that could encode for a coding sequence in M. pneumoniae. For each identifier (column B), we include coordinates information and nucleotide and amino acid length information (columns C-H). Column I includes the gene name when the entry is found annotated in M. pneumoniae. Localization and function are described in columns J and K. Column L includes the operon number in which the annotation would be expressed. We also included transcription-related information average expression (column M; as log2(gene read count/gene length) and estimated average RNA copies per cell (column N) considering 4 RNA sequencing samples covering different growth times (6, 24 and 48 hours, ArrayExpress identifier E‐MTAB‐6203). Column O accounts for the number of mass spectrometry experiments detecting that entry (to a maximum of 116) and column P accounts for the total number of unique tryptic peptides detected. This comes [5], available for 12,426 sequences that present an amino acid length ≥19 (from 116 mass spectrometry experiments, ID PRIDE: PXD008243). Columns Q to T recapitulate protein copies per cell under different conditions (overall, extracting with urea, extracting with SDS and mean, respectively). Column U includes half-lives of the proteins. Columns V and W describe the reference density of insertion and essentiality assigned in previous studies. Column X-AA includes the predicted RanSEPs score, ribosome binding site presence, homology into seven groups: 0—no hits passed the thresholds defined; 1—conserved with an annotated function; 2—conserved as an annotated SEP in NCBI but no associated function; 3—conserved in a different species but target and homologous sequence not found in NCBI; 4—sequence is completely or partially (> 75%) repeated ≥ 3 times in the reference genome; 5—potential pseudogene; and 6—to depict those annotations that are found in the reference NCBI annotation file. , and function expected by homology, respectively. Columns AB to AD cover the output provided by Phobius, including the number of transmembrane segments, presence of signal peptide and transmembrane topology predicted by TM-HMM. Column AE includes the complex information where 1 implies that entry is functional as a monomer, 2 as dimer, and so on. Finally, columns AF-AH will be 1 if the protein is a Lon protease target, a lipoprotein, and/or a truncated gene or pseudogene, respectively, 0 otherwise. Following columns include for every sample presenting selective insertion rates in-frame using the following identifiers separated by underscores: marker (BarnB, Cm or Ery), type (control-AC or selection-BD, antibiotic concentration, sample replicate, frame measured, metric. Metrics account for number insertions in-frame (I), read count (R), linear density from non-coding regions used in the Poisson evaluation (rNC), probability measured (sfNC) and a binary for prediction (pred; 0 - no significant, 1 - significant). Last columns combine the number of samples each annotation has been identified. Notice for barnase library the results need to be interpreted considering it is a negative selection marker inverting the 0 and 1 meaning.