Characterization of missing values in untargeted MS-based metabolomics data and evaluation of missing data handling strategies

Background

Untargeted mass spectrometry (MS)-based metabolomics data often contain missing values that reduce statistical power and can introduce bias in biomedical studies. However, a systematic assessment of the various sources of missing values and strategies to handle these data has received little attention. Missing data can occur systematically, e.g. from run day-dependent effects due to limits of detection (LOD); or it can be random as, for instance, a consequence of sample preparation.

Methods

We investigated patterns of missing data in an MS-based metabolomics experiment of serum samples from the German KORA F4 cohort (n?=?1750). We then evaluated 31 imputation methods in a simulation framework and biologically validated the results by applying all imputation approaches to real metabolomics data. We examined the ability of each method to reconstruct biochemical pathways from data-driven correlation networks, and the ability of the method to increase statistical power while preserving the strength of established metabolic quantitative trait loci.

Results

Run day-dependent LOD-based missing data accounts for most missing values in the metabolomics dataset. Although multiple imputation by chained equations performed well in many scenarios, it is computationally and statistically challenging. K-nearest neighbors (KNN) imputation on observations with variable pre-selection showed robust performance across all evaluation schemes and is computationally more tractable.

Conclusion

Missing data in untargeted MS-based metabolomics data occur for various reasons. Based on our results, we recommend that KNN-based imputation is performed on observations with variable pre-selection since it showed robust results in all evaluation schemes.

     

Biosynthesis of 2-amino-3-hydroxycyclopent-2-enone moiety of bafilomycin in <Emphasis Type="Italic">Kitasatospora cheerisanensis</Emphasis> KCTC2395

Bafilomycins produced by Kitasatospora cheerisanensis KCTC- 2395 belong to the 16-membered macrolactone family plecomacrolide antibiotics. Bafilomycin B₁ contains 2-amino- 3-hydroxycyclopent-2-enone (C₅N), a five membered ring, which gets condensed via an amide linkage to bafilomycin polyketide. To study the biosynthetic pathway of C₅N during bafilomycin biosynthesis in K. cheerisanensis KCTC2395, we attempted the functional analysis of two putative genes, encoding 5-aminolevulinic acid synthase (ALAS) and acyl- CoA ligase (ACL). The amplified putative genes for ALAS and ACL were cloned into the E. coli expression vector pET- 32a(+) plasmid, following which the soluble recombinant ALAS and ACL proteins were purified through nickel-affinity column chromatography. Through HPLC analysis of the enzyme reaction mixture, we confirmed the products of putative ALAS and ACL reaction as 5-aminolevulinic acid (5-ALA) and 5-ALA-CoA, respectively. The optimal pH for the putative ALAS reaction was 7.5, and for putative ACL reaction was 7.0, as confirmed by the colorimetric assay. Furthermore, pyridoxal 5'-phosphate (PLP) was found to be an essential cofactor in the putative ALAS reaction, and ATP was a cofactor for the putative ACL catalysis. Finally, we also confirmed that the simultaneous treatment of putative ACL and putative ALAS enzymes resulted in the production of C5N compound from 5-ALA.     

Discovery of selective urokinase plasminogen activator (uPA) inhibitors as a potential treatment for multiple sclerosis

We report here the design and synthesis of a novel series of benzylamines that are potent and selective inhibitors of uPA with promising oral availability in rat. Further evaluation of one representative (ZK824859) of the new structural class showed that this compound lowered clinical scores when dosed in either acute or chronic mouse EAE models, suggesting that uPA inhibitors of this type could be useful for the treatment of multiple sclerosis.     

Dammarane-type saponins from Gynostemma pentaphyllum

Dammarane-type saponins, gypenosides VN1–VN7 (1–7), were isolated from the total saponin extract of Gynostemma pentaphyllum aerial parts, with their structures elucidated on the basis of spectroscopic and chemical methods. These compounds showed moderate cytotoxic activity against four human cancer cell lines, A549 (lung), HT-29 (colon), MCF-7 (breast), and SK-OV-3 (ovary), with IC₅₀ values ranging from 19.6 ± 1.1 to 43.1 ± 1.0 μM. Regarding the HL-60 (acute promyelocytic leukemia) cell line, compounds 1, 5, and 6 showed weakly active with IC₅₀ values of 62.8 ± 1.9, 72.6 ± 3.6, and 82.4 ± 3.2 nM, respectively, while 2, 3, 4, and 7 were less active with IC₅₀ values >100 μM.     

DNA binding and nucleotide flipping by the human DNA repair protein AGT

O(6)-alkylguanine-DNA alkyltransferase (AGT), or O(6)-methylguanine-DNA methyltransferase (MGMT), prevents mutations and apoptosis resulting from alkylation damage to guanines. AGT irreversibly transfers the alkyl lesion to an active site cysteine in a stoichiometric, direct damage reversal pathway. AGT expression therefore elicits tumor resistance to alkylating chemotherapies, and AGT inhibitors are in clinical trials. We report here structures of human AGT in complex with double-stranded DNA containing the biological substrate O(6)-methylguanine or crosslinked to the mechanistic inhibitor N(1),O(6)-ethanoxanthosine. The prototypical DNA major groove-binding helix-turn-helix (HTH) motif mediates unprecedented minor groove DNA binding. This binding architecture has advantages for DNA repair and nucleotide flipping, and provides a paradigm for HTH interactions in sequence-independent DNA-binding proteins like RecQ and BRCA2. Structural and biochemical results further support an unpredicted role for Tyr114 in nucleotide flipping through phosphate rotation and an efficient kinetic mechanism for locating alkylated bases.     

Repair of oligodeoxyribonucleotides by O(6)-alkylguanine-DNA alkyltransferase

Activity of the DNA repair protein O(6)-alkylguanine-DNA alkyltransferase (AGT) is an important source of tumor cell resistance to alkylating agents. AGT inhibitors may prove useful in enhancing chemotherapy. AGT is inactivated by reacting stoichiometrically with O(6)-benzylguanine (b(6)G), which is currently in clinical trials for this purpose. Short oligodeoxyribonucleotides containing a central b(6)G are more potent inactivators of AGT than b(6)G. We examined whether human AGT could react with oligodeoxyribonucleotides containing multiple b(6)G residues. The single-stranded 7-mer 5'-d[T(b(6)G)(5)G]-3' was an excellent AGT substrate with all five b(6)G adducts repaired although one adduct was repaired much more slowly. The highly b(6)G-resistant Y158H and P140K AGT mutants were also inactivated by 5'-d[T(b(6)G)(5)G]-3'. Studies with 7-mers containing a single b(6)G adduct showed that 5'-d[TGGGG(b(6)G)G]-3' was more poorly repaired by wild-type AGT than 5'-d[T(b(6)G)GGGGG]-3' and 5'-d[TGG(b(6)G)GGG]-3' and was even less repairable by mutants Y158H and P140K. This positional effect was unaffected by interchanging the terminal 5'- or 3'-nucleotides and was also observed with single-stranded 16-mer oligodeoxyribonucleotides containing O(6)-methylguanine, where a minimum of four nucleotides 3' to the lesion was required for the most efficient repair. Annealing with the reverse complementary strands to produce double-stranded substrates increased the ability of AGT to repair adducts at all positions except at positions 2 and 15. Our results suggest that AGT recognizes the polarity of single-stranded DNA, with the best substrates having an adduct adjacent to the 5'-terminal residue. These findings will aid in designing novel AGT inhibitors that incorporate O(6)-alkylguanine adducts in oligodeoxyribonucleotide contexts.