Fix: GPR curation for acyl-CoA hydrolysis and nicotinate and nicotinamide metabolism

[JHL-452b]

Main improvements in this PR:

As proposed in #890

• Remove ENSG00000112304 from MAR00242;
• Replace ENSG00000205669 with ENSG00000159445 from MAR00324;
• Replace ENSG00000205669 with ENSG00000159445 from MAR00332;
• Add ENSG00000159445 for MAR00354;
• Replace ENSG00000177465 and ENSG00000101473 with ENSG00000172497 and ENSG00000159445 from MAR03009;
• Remove ENSG00000177465 from MAR03009;
• Remove ENSG00000112304 from MAR00210;
• Replace ENSG00000205669 with ENSG00000159445 from MAR00426;
• Replace ENSG00000205669 with ENSG00000159445 from MAR00438;
• Remove ENSG00000205309, ENSG00000076685 from MAR04252;
• Remove ENSG00000205309, ENSG00000076685 from MAR04264;
• Remove ENSG00000162836 from MAR04270;
• Remove ENSG00000198805 from MAR04662;
• Remove ENSG00000109743 from MAR08788.

I hereby confirm that I have:

• Tested my code on my own computer for running the model
• Selected develop as a target branch
• Any removed reactions and metabolites have been moved to the corresponding deprecated identifier lists

[github-actions]

This PR has been automatically tested with GH Actions. Here is the output of the MACAW test:

Starting dead-end test...
 - Found 1383 dead-end metabolites.
 - Found 1139 reactions incapable of sustaining steady-state fluxes in either direction due to these dead-ends.
 - Found 1368 reversible reactions that can only carry steady-state fluxes in a single direction due to dead-ends.
Starting duplicate test...
 - Skipping redox duplicates because no redox_pairs and/or proton_ids were provided.
 - Found 379 reactions that were some type of duplicate:
   - 0 were completely identical to at least one other reaction.
   - 13 involve the same metabolites but go in the opposite direction or have the opposite reversibility as at least one other reaction.
   - 379 involve the same metabolites but with different coefficients as at least one other reaction.

This and a more detailed output from MACAW are also committed to data/macawResults/.

Note: In the case of multiple test runs, this post will be edited.

[JHL-452b]

I'm sorry I didn't check carefully enough before. I am now realizing that the modification I proposed for MAR03009 may have a problem.

What we could know here is that ENSG00000172497 and ENSG00000159445 have a hydrolyzing function for most acyl-CoAs in the cytosol. However, MAR03009 is located in peroxisomal rather than cytosol. This is the root cause of the problem.

The original GPR of MAR03009 is ENSG00000101473 or ENSG00000177465. Both of them have the hydrolyzing function for acyl-CoAs and located in peroxisomal. So, I think MAR03009 should keep the original GPR. What do you think? @Devlin-Moyer

[Devlin-Moyer]

Ah I see that I mistakenly thought that MAR03009 involved cytosolic metabolites and not peroxisomal ones when writing my earlier message. After reassessing the situation with that in mind:

• ACOT8 (ENSG00000101473) does belong in the GPR of MAR03009, so it doesn't need to be removed
• ACOT4 (ENSG00000177465) should still be removed, because, while it is capable of hydrolyzing acyl-CoAs in peroxisomes, it is not active towards propanoyl-CoA (see figure 4A of this paper)
• THEM4 (ENSG00000159445) should not be added, as I had previously suggested, as it is not localized to peroxisomes
• ACOT12 (ENSG00000172497) might also be appropriate to include in the GPR of MAR03009 since it is known to be active against propanoyl-CoA (source), but that same paper also says that "only a minor fraction of the protein associates with peroxisomes", so it might be fine to leave it out

To summarize: I now think the GPR of MAR03009 should be either ENSG00000101473 or ENSG00000101473 or ENSG00000172497, depending on how we feel about the apparently "minor fraction" of peroxisomal ACOT12

[JHL-452b]

ACOT4 (ENSG00000177465) should still be removed, because, while it is capable of hydrolyzing acyl-CoAs in peroxisomes, it is not active towards propanoyl-CoA (see figure 4A of this paper)

The human ACOT4(ENSG00000177465) possesses activity with medium- to long chain acyl-CoAs in addition to the activity with succinyl-CoA and glutaryl-CoA. But the corresponding C4-monocarboxy-lyl-CoA (butyryl-CoA) was not a substrate, and in contrast the C12 dicarboxylyl-CoA was not a substrate(even though the corresponding C12 monocarboxylyl-CoA was a good substrate). Therefore, I believe that ACOT4 is also inactive towards propanoyl-CoA (similar to butyryl-CoA) in MAR03009. Agree to remove ACOT4 (ENSG00000177465) for MAR03009.

As for ACOT12(ENSG00000172497), I think we should not add it to the GPR of MAR03009 for the time being. We will make a judgment after finding stronger evidence.

[github-actions]

This PR has been automatically tested with GH Actions. Here is the output of the gene essentiality test:

     cellLine     TP     TN     FP    FN     accuracy    sensitivity    specificity       F1         MCC   
    __________    __    ____    __    ___    ________    ___________    ___________    ________    ________

    {'DLD1'  }    37    2165    62    277    0.86659       0.11783        0.97216       0.17918     0.15306
    {'GBM'   }    34    2147    65    294    0.85866       0.10366        0.97061       0.15925     0.12869
    {'HCT116'}    46    2189    57    305    0.86061       0.13105        0.97462       0.20264     0.18513
    {'HELA'  }    31    2246    72    249    0.87644       0.11071        0.96894       0.16188     0.12659
    {'RPE1'  }    15    2186    84    255    0.86654      0.055556          0.963      0.081301    0.029544
    {'all'   }     7    2386    96    109    0.92109      0.060345        0.96132      0.063927    0.022933

Note: In the case of multiple test runs, this post will be edited.

[feiranl]

@edkerk Could check this PR?

[edkerk]

LGTM