Table 1 Properties of reanalyzed studies

From: Community benchmarking and evaluation of human unannotated microprotein detection by mass spectrometry based proteomics

Citation	sORF database size	Considered noncanonical PSMs	Reported sORFs with MS support	HLA or non-HLA	Source material	Annotation definition	Reported false discovery rate
Cao et al. ⁴⁴	Three-frame translation of transcriptome	28	17	non-HLA	HEK293T	Human UniProtKB 2019	1% at peptide and protein level, proteome-wide
Bogaert et al. ²⁸	16,919	8	6	non-HLA	HEK293T cellular cytosol	Human UniProtKB/Swiss-Prot 2021	<1% peptide, <2.5% protein, proteome-wide
Chothani et al. ⁴	7767	5763	614	non-HLA	NHDF and HUVEC (Slany et al. ⁶⁶.), ES (Shekari et al. ⁶⁷.), Heart (Doll et al. ⁶⁸.)	Human UniProtKB 2017	1% PSM level, unannotated specific
Duffy et al. ³⁴.	38,187	2445	366	non-HLA	Adult brain, Prenatal brain, hESC-derived neurons	Human UniProtKB	1% at peptide and protein level, proteome-wide
Douka et al. ⁴⁰.	45	18	8	non-HLA	SH-SY5Y cells (Murillo et al. ⁶⁹. and Brenig et al. ⁷⁰.)	Human UniProtKB 2019	10% at peptide level, proteome-wide
Prensner et al. ³³.	553	6236	140	HLA and non-HLA	14 published mass spectrometry datasets	UCSC RefSeq	1% at PSM level, proteome-wide
Ouspenskaia et al. ²⁹.	237,437	9985	4903	HLA and non-HLA^*	Lymphoblastoid cell line (Sarkizova et al. ⁷¹.), patient-derived melanoma cell line, patient-derived glioblastoma cell line (Shraibman et al. ⁷².), chronic lymphocytic leukemia tumor, ovarian carcinoma, renal cell carcinoma	Annotated genes on UCSC Genome Browser hg19	1% at PSM level, class-specific FDR for each type of unannotated ORF (e.g., uORF, dORF)
Chen et al. ³⁵.	7824	33	12	HLA and non-HLA†	iPSCs	Human UniProtKB	1% at PSM level, proteome-wide
Chong et al. ³².	Three-frame translation of transcriptome	2597	384	HLA	Patient-derived melanoma cell lines and lung cancer samples with matched normal tissues	Human UniProtKB/TrEMBL 2018	Class-specific FDR for unannotated, keep only PSMs identified by both Comet and MaxQuant. Estimated FDR < 0.001%
Martinez et al. ⁴².	7554	1160	319	HLA	Six cancer cell lines from Bassani-Sternberg 2015 (25576301): B-cells EBV transformed, B-cell leukemia, basal like breast cancer, colon carcinoma, primary fibroblast	Human UniProtKB/Swiss-Prot	1% FDR at peptide level, proteome-wide
van Heesch et al. ⁶.	1598	1942	500	non-HLA	Heart (Doll et al., 29133944), iPSC-derived cardiomyocytes	Human UniProtKB 2017	1% targeted FDR, 50-60% estimated FDR
Lu et al. ⁵⁹.	2969	964	308	non-HLA	Cell lines: lung, colorectal cancer, liver cancer, cervical cancer	Human UniProtKB/Swiss-Prot	1% FDR at PSM, peptide and protein level.

^*Only HLA spectra were evaluated. †Only non-HLA spectra were evaluated.
List of all studies reanalyzed. sORF database size indicates the number of sORFs in the protein sequence database in the MS analysis for each study. The number of these ORFs with proteomic support according to the study is also given. Considered noncanonical PSMs is the number of PSMs supporting a sORF-encoded protein reported in each study for which we could obtain the necessary information to evaluate; PSMs actually evaluated were selected randomly from this set. Annotation definition indicates the database used by each study to define the set of annotated or “canonical” proteins; all other proteins are considered to be unannotated, sORF-expressed proteins. Reported false discovery rate indicates the FDR given in each study for the list of sORF detections and whether this was calculated proteome-wide (a common FDR considering both unannotated and annotated proteins) or specific to the unannotated proteins.

Back to article page

Table 1 Properties of reanalyzed studies

Search

Quick links