{"id":"W6949309693","doi":"10.5281/zenodo.14014029","title":"A strategy for scalable data collection of soluble protein expression in diverse hosts","year":2024,"lang":"en","type":"article","venue":"Zenodo (CERN European Organization for Nuclear Research)","topic":"vaccines and immunoinformatics approaches","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute of Genetics","funders":"","keywords":"Protein expression; Expression (computer science); Inference; Scalability; Field (mathematics); Biomanufacturing; Data collection; Information repository","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004284818,0.00007517415,0.00008037846,0.0001004328,0.0003036884,0.0001935685,0.0004790471,0.00006091844,0.0003147522],"category_scores_gemma":[0.0001725596,0.00007252116,0.000026244,0.0002321919,0.00002815589,0.00003795379,0.0008324923,0.00007548137,0.0000734885],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002349072,"about_ca_system_score_gemma":0.000009655173,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001321128,"about_ca_topic_score_gemma":0.000001006935,"domain_scores_codex":[0.9992533,0.00005414956,0.0001873401,0.0002407618,0.000100569,0.0001638721],"domain_scores_gemma":[0.9994055,0.000004328533,0.00004687566,0.0003670876,0.0001401913,0.00003598604],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002896251,0.0001656954,0.000007150676,0.0005481206,0.00004919657,0.000001669445,0.0003765622,0.0002721355,0.7728882,0.0005324858,0.178187,0.04668214],"study_design_scores_gemma":[0.0008108722,0.0006444079,0.0001501447,0.0001377712,0.00001124525,0.00002081286,0.0005208307,0.02091112,0.1093745,0.000114544,0.8671394,0.0001643494],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8984262,0.002107002,0.05330093,0.0005678638,0.0003405441,0.003940611,0.001121084,0.0003600798,0.03983574],"genre_scores_gemma":[0.9928102,0.0001190467,0.001076888,0.000009292816,0.00006055566,2.898649e-7,0.004455313,0.0003002082,0.001168194],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6889524,"threshold_uncertainty_score":0.3446316,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05812364569814416,"score_gpt":0.2714072730242544,"score_spread":0.2132836273261102,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}