Not a person — a BERIL service account. Listed as an author on the projects below.
Contributions
The Acinetobacter baylyi ADP1 Data Explorer project analyzed a user-provided SQLite database for Acinetobacter baylyi ADP1 and 13 related genomes, and found that it holds 15 tables, 461,522 total rows and 135 MB of data. 1 The project found that the central genome_features table contains 5,852 genes and 51 annotation columns spanning six data modalities. 1 These modalities were TnSeq (transposon-insertion sequencing) essentiality with 58% coverage, FBA (flux balance analysis, constraint-based prediction of metabolic fluxes) with 15% coverage, mutant growth fitness on 8 carbon sources with 39% coverage, proteomics across 7 strains with 41% coverage, and pangenome classification with 54% coverage. 1 The sixth modality was functional annotation through COG (Clusters of Orthologous Groups), KEGG KO, Pfam and UniRef, with 34-55% coverage. 1 No single gene has data across all six modalities, although the project found substantial pairwise overlaps among essentiality, pangenome and proteomics. 1 The project documented 7 engineered ADP1 strains, each with proteomics data. 1 These were wild-type ADP1 plus 6 derivatives with aromatic amino acid pathway modifications involving ΔaroF and ΔaroG or dgoA variants. 1 Protein abundance was measured for 2,383 genes, and the high cross-strain correlation indicated targeted rather than global effects of the modifications. 1
The project tested five connection types against the KBase Data Lakehouse and found that four matched at greater than 90%. 1
- 13 of 13 genome IDs matched the pangenome. 1
- 1,210 of 1,330 reactions (91%) matched biochemistry. 1
- 230 of 230 compounds matched biochemistry. 1
- 4,891 of 4,891 cluster IDs matched the pangenome through mapping. 1
ADP1 had 0 matches among 1 Fitness Browser organism query and was absent from the Fitness Browser. 1 The project reported that the 120 unmatched reactions may be custom or draft reactions not yet in ModelSEED. 1 Because ADP1-style and KBase Data Lakehouse cluster IDs had 0% direct string match, the project built a bridge through the gene_genecluster_junction table. 1 All 4,891 KBase Data Lakehouse clusters mapped to 4,081 unique ADP1 clusters, a 100% gene-level match across 43,754 genes. 1 The project cautioned that this mapping passes through 3 tables, and that the indirection could introduce edge cases where clusters split or merge. 1
The project compared 866 genes that had both FBA predictions and TnSeq calls. 1 It found 639 (73.8%) concordant and 227 discordant, and named the discordant genes as candidates for model refinement or as possible regulatory effects not represented by FBA. 1 The project found 499 genes essential on minimal media versus 346 on LB, and FBA flux classes changed between rich and minimal media for 177 of 866 genes (20%). 1 Essential genes were more annotation-rich than dispensable genes. 1 COG assignments covered 33% of essential genes versus 5% of dispensable genes, and KO assignments covered 92% versus 53%. 1 Essential genes were also more likely to be core pangenome genes, and the roughly 8% of essential genes lacking KO assignments were flagged as potential novel essential functions. 1 Mutant fitness across 8 carbon sources had a mean pairwise correlation of 0.44. 1 Urea was nearly uncorrelated with quinate (r = 0.11) and weakly correlated with all other conditions (r = 0.12-0.28). 1 Butanediol-acetate and butanediol-lactate showed the strongest correlations (r = 0.58 and r = 0.53). 1 The project noted that, because ADP1 is absent from the Fitness Browser, these fitness data are a resource not otherwise available in the KBase Data Lakehouse. 1 The project slots these results into gene-essentiality and condition-specific-fitness. 1
On metabolic reactions, the project found that 1,248 of 1,330 unique reactions (94%) were core across all 14 genomes. 1 Another 62 were variable and 20 were genome-unique. 1 Gapfilling accounted for 7.7% of reactions on average, with 243 missing functions cataloged. 1 Of 121,519 growth phenotype predictions, 105,376 (87%) required at least one gapfilled reaction, and false negatives had higher mean gap counts than correct predictions. 1 The project stated that this dependence limits interpretation of the predictions. 1 The project noted that the FBA-TnSeq concordance analysis is limited to 866 genes and that the database covers only A. baylyi. 1 It presented the discordant-gene enrichment tests and cross-species comparisons as future analyses rather than completed findings. 1 The project slots these results into metabolic-model-gapfilling. 1