Human 1.19

[JHL-452b]

Main improvements in this PR:

• Fixes(PR chore: update actions #818 )

  • resolve the Node 16 warning as discussed in bug in merge check-yaml validation warning report #817

• Fixes(PR fix: modified unbalance reactions #816 )

  • Refine unbalance reactipns as discussed in Unbalance reactions checked in the new content of Human-GEM #814 and Unbalance reactions about FAD/Oxygen #813
    • change the charge of mets MAM20004n, MAM20006c, MAM20018r, MAM20021i, MAM20024r, MAM20025i, MAM20026r, MAM20027i, MAM20028r, MAM20029i, MAM20030r, MAM20031i, MAM20041r, MAM20042c, MAM20043r, MAM20044c, MAM20056r, MAM20057c from 0 to -1
    • remove MAM02039e in MAR01444; remove MAM02039i in MAR20022, MAR20024, MAR20026, MAR20028 and MAR20030; remove MAM02039r in MAR20023, MAR20025, MAR20027, MAR20029 and MAR20021;
    • add MAM02039m in MAR20111 as a product; add MAM01422m in MAR20112 as a product; and add MAM02039c in MAR03481 as a product;
    • replace the formula of MAM01657r to "C40H65O7P2"; replace the formula of MAM20082m from None to "C11H23N1O2S2"; replace MAM02041x with MAM01803x in MAR08415; replace MAM02630x with MAM01802x in MAR08415; replace MAM02041x with MAM01803x in MAR03984; replace MAM02630x with MAM01802x in MAR03984; replace MAM02041x with MAM01803x in MAR12372; replace MAM02630x with MAM01802x in MAR12372

• Fixes(PR Fix Mitochondrial Thiamine Transport Reactions #815 )

  • Fixes the transport of thiamine Di- and Monophosphate between cytosol and mitochondria as discussed in Transport of Thiamine Di- and Monophosphate Between Cytosol and Mitochondria #783
    • Created MAR20181 with ENSG00000125454 as GPR;
    • Created MAR20182 with ENSG00000125454 as GPR;
    • Created MAR20813 with ENSG00000125454 as GPR;
    • Removed MAR01788, MAR01789, MAR04205, MAR08745, and MAR08747 for lacking evidence and being redundant with the three new reactions above.

• Fixes(PR fix: standardize all names of exchange reactions #811）

  • Standardize all names of exchange reactions as discussed in missing reaction names #181

• Fixes(PR fix: GPRs for mitochondrial 2-enoyl-CoA hydratase reactions #808)

  • Refine GPRs of Mitochondrial 2-enoyl-CoA Hydration Reactions as explained in fix: GPRs for mitochondrial 2-enoyl-CoA hydratase reactions #808
    • Set GPRs of MAR01163, MAR01177, MAR01182, MAR01206, MAR01219, MAR01224, MAR01255, MAR01272, MAR03108, MAR03112, MAR03116, MAR03122, MAR03129, MAR03171, MAR03175, MAR03179, MAR03183, MAR03187, MAR03219, MAR03223, MAR03227, MAR03231, MAR03235, MAR03241, MAR03245, MAR03252, MAR03260, MAR03277, MAR03281, MAR03285, MAR03449, and MAR03453 to ENSG00000084754 (HADHA)
    • Set GPRs of MAR03136, MAR03143, MAR03150, MAR03157, MAR03191, MAR03195, MAR03199, MAR03290, MAR03523, and MAR03528 to ENSG00000084754 or ENSG00000127884 (HADHA or ECHS1)
    • Set GPRs of MAR03164, MAR03203, MAR03532, MAR03753, MAR03785, and MAR04735 to ENSG00000127884 (ECHS1)

• Fixes(PR Remove Tagatose-1,6-Bisphosphate #807)

  • Removes MAR04774, MAR04775, and MAR02955c as they seem could not occur in human cells;
  • Updates the references of MAR09417, MAR08760, MAR01472, MAR01038, MAR08761 and MAR08762 as discussed in Tagatose is probably not metabolized by human cells #754

• Fixes(PR Fix Stereochemistry of Phytanic Acid Oxidation #805)

  • Curate stereochemistry of phytanic acid metabolism as discussed in Stereochemistry of Phytanic Acid Metabolism #773
    • Rename MAM02746e, MAM02746c and MAM02746x to "(3R)-phytanic acid" and MAM02747c and MAM02747x to "(3R)-phytanoyl-CoA"
    • Brought back previously-deprecated MAM03884e, MAM03884c and MAM03884x and renamed them to "(3S)-phytanic acid"
    • Brought back previously-deprecated MAR00659 and MAR01617 (transport reactions for MAM03884)
    • Replace MAM02746c with MAM03884c in MAR03388
    • Change MAR03481 to be MAM02747c + MAM01371c + 2 MAM02040c -> MAM02746x + MAM01597c + MAM02039c + MAM01285c + MAM02751c, GPR: ENSG00000117528, references: PMID:24333844;PMID:33500543
    • Change MAR06779 to be MAM03362c + MAM01371c + 2 MAM02040c -> MAM03384x + MAM01597c + MAM02039c + MAM01285c + MAM02751c, GPR: ENSG00000117528, references: PMID:24333844;PMID:33500543
    • Replace MAM02746x with MAM03884x in MAR03389
    • Rename MAM00655x to "(3R)-2-hydroxyphytanoyl-CoA"
    • Create MAM20083x to represent peroxisomal (2R)-pristanal, and replace MAM00564x ((2S)-pristanal) with MAM20083x in MAR03486
    • Rename MAM02766e, MAM02766c, and MAM02766x to "(2S)-pristanic acid"
    • Brought back previously-deprecated MAR03488 and MAM00077x
    • Replace MAM00564x with MAM20083x in MAR03488, and replace MAM02766x with MAM00077x in MAR03493
    • Create MAM00077c and MAM00077e to represent cytosolic and extracellular (2R)-pristanic acid, and create MAR20179 to represent export of (2R)-pristanic acid from cytosol to extracellular environment
    • Replace MAM02766c with MAM00077c in MAR03390
    • Create MAR20180 to represent exchange of (2R)-pristanic acid

• Fixes(PR Remove MAR00965 and Fix GPR of MAR00970 #794)

  • Removes MAR00965 and fixes the gpr of MAR00970 as explained in Incorrect Beta-Oxidation Reaction for (4Z,7Z,10Z,13Z,16Z)-docosapentaenoyl-CoA #748

• Fixes(PR Merge Duplicate Mitochondrial Linoleic and Gamma-Linolenic Acid Beta-Oxidation Pathways #793 )

  • Curate duplicate Beta-oxidation pathways for Linoleoyl- and gamma-Linolenoyl-CoA as explained in Merge Duplicate Mitochondrial Linoleic and Gamma-Linolenic Acid Beta-Oxidation Pathways #793
    • Removes MAR05358 for being a duplicate of MAR03452 + MAR03453 + MAR03454 + MAR03455
    • Removes MAR05195 for being a duplicate of MAR03456
    • Removes MAR05302 for being a duplicate of MAR03457
    • Removes MAR05229 for being a duplicate of MAR03458
    • Removes MAR05208 for being a duplicate of MAR03281 + MAR03282 + MAR03283
    • Removes MAR05356 for being a duplicate of MAR03275 + MAR03277 + MAR03278 + MAR03279
    • Removes MAR05191 for being a duplicate of MAR03280 + MAR03281 + MAR03282 + MAR03283
    • Removes MAR05123 for being a duplicate of MAR03284 + MAR03285 + MAR03286 + MAR03287
    • Removes MAR05075 for being a duplicate of MAR03288
    • Removes MAR05072 for being a duplicate of MAR03290 + MAR03292 + MAR03293
    • Replaces MAM03274m with MAM00091m in MAR05360
    • Removes MAM03274m for being a duplicate of MAM00091m
    • Removes MAM03653m for being a duplicate of MAM00072m
    • Removes MAM03221m for being a duplicate of MAM00089m
    • Removes MAM03182m for being a duplicate of MAM03009m
    • Removes MAM03978m for being a duplicate of MAM01576m
    • Removes MAM03203m for being a duplicate of MAM01575m
    • Removes MAM02698m for being a duplicate of MAM03021m
    • Removes MAM03650m for being a duplicate of MAM01577m

• Fixes(PR Resolve Duplicate Phytanoyl-CoA Alpha-Oxidation Reactions #792)

  • Removes MAR05187 for being a duplicate of MAR03486 + MAR03386;
  • Removes MAR05186 for being a misrepresentation of ALDH3A2;
  • Set GPR of MAR03386 to ENSG00000072210;
  • Removes ENSG00000182591 from the GPR of MAR03486 and remove it from Human-GEM. All these are explained in Duplicate Reactions For Alpha-Oxidation of Phytanoyl-CoA #775

• Fixes(PR Remove EHHADH from the GPR of MAR03288 #790)

  • removes EHHADH (ENSG00000113790) from the GPR of MAR03288 as discussed in Bad GPRs for Mitochondrial delta-3-delta-2-enoyl-CoA Isomerization Reactions #756

• Fixes(PR Remove Duplicate Decanoyl-CoA to Octanoyl-CoA Beta-Oxidation Reactions #789)

  • Removes MAR05019 for being a duplicate of MAR03142 + MAR03143;
  • Removes MAR04967 for being a duplicate of MAR03143 + MAR03144 + MAR03146;
  • Removes MAR05024 for being a duplicate of MAR03142 + MAR03143 + MAR03144 + MAR03146. All these are discussed in Duplicate Beta-oxidation Reactions for Decanoyl-CoA #745

• Fixes(PR fix: revise rxns and GPRs associated with DECR1 and DECR2 #788)

  • Remove MAR03322 and MAR03296 due to cofactor mismatch and remove MAR04005, and associated MAM00678c and MAM03035c, due to compartment mismatch as reported in DECR1 and DECR2 use NADPH, not NADH #736;
  • Set the GPRs of MAR01176, MAR01181, MAR01218, MAR01222, MAR03446, and MAR03457 to ENSG00000104325 and set the GPRs of MAR01297, MAR03461, and MAR03471 to ENSG00000242612 as reported in Localization of DECR1 and DECR2 #755

• Fixes(PR fix: remove genes only expressed in peroxisome from beta-oxidation reactions in mitochondria #770)

  • Remove ENSG00000115425 from MAR02028, MAR02148, MAR02149, MAR02183, MAR02187 and MAR02207 as explained in Genes expressed in peroxisome but found in GPR of beta-oxidation reactions in mitochondria #767

• Fixes(PR Remove Duplicate Linolenoyl-CoA Beta-Oxidation Reactions #768)

  • remove inaccurate reaction MAR00789 and duplicate reaction MAR03426 of beta-oxidation of linolenoyl-CoA as explained in Duplicate Beta-Oxidation Pathways for Linolenoyl-CoA #765

• Fixes(PR Fix GPRs for MAR03269 and MAR03380 #760)

  • Refine GPR of MAR03269 and MAR03380 to ENSG00000084754 or ENSG00000072506 genes, as discussed in GPRs for Mitochondrial 3-Hydroxyacyl-CoA Oxidation Reactions #758

• Fixes(PR Fix inconsistent representation of thiamine and lipoic acid in 2-oxoacid dehydrogenase reactions #753)

  • revise inconsistent representation of Thiamine in 2-Oxoacid Dehydrogenase Complexes as explained in Inconsistent Representation of Thiamine's Role in 2-Oxoacid Dehydrogenase Complexes #744
    • Creates MAM20077m to represent 2-(4-carboxy-1-hydroxybutyl)thiamine diphosphate and MAR20173 with ENSG00000105953 or ENSG00000181192 as GPR
    • Replaces MAM00670m with MAM20077m, replaces MAM01596m with MAM02984m, and removes MAM02039m in MAR04599
    • Creates MAM20078m to represent 2-(2-methyl-1-hydroxypropyl)thiamine diphosphate and MAR20174 with ENSG00000248098 and ENSG00000083123 as GPR
    • Replaces MAM00824m with MAM20078m, replaces MAM01596m with MAM02984m, and removes MAM02039m in MAR06416
    • Creates MAM20079m to represent 2-(2-methyl-1-hydroxybutyl)thiamine diphosphate and MAR20175 with ENSG00000248098 and ENSG00000083123 as GPR
    • Replaces MAM00669m with MAM20079m, replaces MAM01596m with MAM02984m, and removes MAM02039m in MAR06419
    • Creates MAM20080m to represent 2-(3-methyl-1-hydroxybutyl)thiamine diphosphate and MAR02176 with ENSG00000248098 and ENSG00000083123 as GPR
    • Replaces MAM01013m with MAM20080m, replaces MAM01596m with MAM02984m, and removes MAR02039m in MAR06421
    • Creates MAM20081m to represent 2-(1-hydroxypropyl)thiamine diphosphate and MAM20082m to represent S-propanoyldihydrolipoamide
    • Changes MAR03800 to: MAM00671m + MAM02984m + MAM02039m -> MAM20081m + MAM01596m, GPR: ENSG00000248098 and ENSG00000083123
    • Creates MAR20177 with ENSG00000248098 and ENSG00000083123 as GPR and MAR20178 with ENSG00000137992 as GPR.

• Fixes(PR Remove MAR03422 #752)

  • Remove duplicate Beta-oxidation reaction MAR03422 as explained in Duplicate Beta-Oxidation Reactions for (4Z,7Z,10Z,13Z,16Z)-docosapentaenoyl-CoA  #746

• Fixes(PR Fix Duplicate Octanoyl-CoA Beta-Oxidation Reactions #751)

  • Remove duplicate octanoyl-CoA Beta-oxidation reactions, include MAR06329, MAR06283, MAR06279, MAR05449 and MAR01009 as explained in Overlapping duplicate reactions for beta-oxidation of octanoyl-CoA #747

• Fixes(PR Fix BCKDH Genes and Remove Mito & Nuc Thiocyanate <-> Hypothiocyanite Reactions #750)

  • Refine gpr involving BCKDHA (ENSG00000248098), BCKDHB (ENSG00000083123), and TMEM91 (ENSG00000142046) genes, as explained in Genes Associated with the BCKDH Complex and Thiocyanate #742
    • Remove MAR02532 and MAR02534;
    • Remove MAM02158m, MAM02158n, and MAM02986n;
    • Remove ENSG00000130508 and ENSG00000117592 from the GPR of MAR06523;
    • Add ENSG00000137992 to the GPRs of MAR06416, MAR06419, and MAR06421;
    • Add PMID:11013238 and PMID:31989833 to the references of MAR06523;
    • Removes ENSG00000142046 from the GPRs of MAR06416, MAR06419, MAR06421, and MAR06523 as well as the list of genes in Human-GEM

• Fixes(PR Fix Duplicate BCKDH Reactions #743)

  • Remove duplicate branched-chain alpha-ketoacid dehydrogenase reaction MAR03748 and MAR03780 as explained in Duplicate Branched-Chain Ketoacid Dehydrogenase Reactions #741

• Features (PR Feat: added smiles and inchi of metabolites #731)

  • Add SMILES and InChI of metabolites in Human-GEM as discussed in Feat: a task for SMILES of metabolite should be included #728

• Feature(PR feat: add gene essentiality check using DepMap #574)

  • Add evaluation of gene essentiality using DepMap data.

[feiranl]

Thanks Ed for fixing the issue!
Could you paste the paste the Essentiality here? @JHL-452b

[edkerk]

There were two problems:

• One of the earlier PRs was made from a fork with a branch that was derived from an earlier develop branch.
• The YAML file was not fully sorted alphabetically.

The latter could be resolved by a round of import-export. The resulting model could then be diff'ed, and the differences to main were manually checked to correspond the original PRs.

The first problem could only be resolved by a -Xtheirs-merge to main, which then closes the PR.

To reduce the risk of similar problems in the future:

• Do not allow PRs made from forks, only from feature branches in the SysBioChalmers/Human-GEM repo. In forks it is less obvious that the feature branch was based on the most recent develop branch. Regular contributors should be given write access to make branches (as is already customary).
  • If a PR is made from a fork (by an external collaborator), then a new feature branch in SysBioChalmers/Human-GEM should be made, and all PR commits cherry-picked to that branch.
• Before merging a PR, go through a round of import-export (with importYaml -> exportYaml) to ensure that the YAML file is (alphabetically) ordered.

[JHL-452b]

Updated essentiality evaluation using combined (all) Hart2015 datasets:

version	TP	TN	FP	FN	accuracy	sensitivity	specificity	F1	MCC
v1.12	40	2333	175	77	0.904000000000000	0.341880341880342	0.930223285486443	0.240963855421687	0.204768560393159
v1.13	40	2334	174	77	0.904380952380952	0.341880341880342	0.930622009569378	0.241691842900302	0.205504558241293
v1.14	40	2334	175	77	0.904036557501904	0.341880341880342	0.930251096054205	0.240963855421687	0.204787829413435
v1.15	40	2342	168	77	0.906737723639132	0.341880341880342	0.933067729083665	0.246153846153846	0.210053956889428
v1.16	41	2308	202	76	0.89417587	0.35042735	0.91952191	0.22777778	0.19220101
v1.17	41	2263	237	75	0.880733944954129	0.353448275862069	0.905200000000000	0.208121827411168	0.172768250444715
v1.18	42	2260	245	74	0.878290728729493	0.362068965517241	0.902195608782435	0.208436724565757	0.174052567331894
v1.19	38	2371	131	78	0.920168067226891	0.327586206896552	0.947641886490807	0.266666666666667	0.230477038058642

Compared with Human 1.18, Human 1.19 seems to have changed a lot, especially the TN has increased significantly.

The organ-specific models were generated by the ftINIT (1+1 mode) , and the .code/evaluateHart2015Essentiality function was used to obtain the gene evaluation results.

[johan-gson]

Hmm, and everything was done exactly the same for all of these versions? Would be good to figure out what makes the difference? Is it possible to do interval halving, i.e., include half of the changes and see if it changed then, then close in on the interval by taking/removing half of the changes where the change occured? For the Hart dataset this may not be that heavy?

[JonathanRob]

Actually the drop in FP is the biggest relative change; it seems overall to have improved a decent amount, though at the expense of a bit lower sensitivity - a good trade off in my opinion.

[JHL-452b]

Hmm, and everything was done exactly the same for all of these versions? Would be good to figure out what makes the difference?

I'm not sure what was done for previous versions. But we do need to figure out what's causing the difference.

Is it possible to do interval halving, i.e., include half of the changes and see if it changed then, then close in on the interval by taking/removing half of the changes where the change occured?

Do you mean to try to impose half of changes to the model and evaluate it, if I understand correctly. This sounds like a good idea. @johan-gson

[johan-gson]

Yep, and then depending on in which half the changes, split that in half and include all changes up to that point, and os on. Will give log(n) complexity :)

[mihai-sysbio]

Fantastic to have this merged! And indeed a positive (great!) suprise with the increase in the MCC. One could use also git bisect, to avoid playing with commits manually.

[mihai-sysbio]

.code/evaluateHart2015Essentiality function was used to obtain the gene evaluation results.

@JHL-452b there is a (long-standing) PR to run this function automatically on all pull requests to main and develop. Would you want to have a look in #675?

[JHL-452b]

@JHL-452b there is a (long-standing) PR to run this function automatically on all pull requests to main and develop. Would you want to have a look in #675?

Oh yeah, it truly helps a lot. I will have a try in the next few days. Thanks！