Among water pollutants, synthetic dyes and heavy metals pose severe health threats and the latter is currently a concern in many parts of Papua New Guinea (PNG) especially in urban and mining areas  . Many reviews are dedicated to identifying and presenting low-cost sorbents for heavy metal removal and recovery  . Therefore it is important to find an economically viable treatment method for drinking water as well as wastewater to minimize the concentrations of heavy metals in water bodies. The objective of this research was to investigate the effectiveness of the root husk of cassava (Manhiotesculenta) to remove Pb(II) and Cu(II) ions from solutions through the process of bioadsorption. Among several techniques, adsorption using solid materials, such as modified silica, alumina, activated carbon, and resin, is considered as the superior methods for metal ion extraction process   . However, these materials are expensive to acquire in many developing countries  including PNG. According to McKay , the criteria for adsorption as an efficient treatment process is to first search for an adsorbent with high selectivity, high adsorbent capacity, abundant availability at low cost. Bioadsorption using agricultural wastes have the advantage of being readily available in tones and can bind metals through various mechanisms such as chelation/coordination, ion exchange, precipitation, and sorption by other physical and chemical forces . Cassava is a common food crop in many countries around the world and managing and dumping of their root husks become a burden so they can be utilized effectively for purifying water as well as for trapping valuable metals.
FTIR is an important tool used in identifying the functional groups available on biomaterials which are responsible for the heavy metal ions coordination. FTIR uses an interferometer and flourier transformations to obtain a spectrum corresponding to wave numbers that can be assigned to certain functional groups of a biomaterial. Elemental analysis (CHNS) was used to determine the percentage elemental composition of carbon, hydrogen, nitrogen, and sulphur. This provides the basis for identifying the species of elements that can either act as Lewis base and anions for metal sequestration. Bioadsorption studies are usually carried out in batch processes as a function of experimental perimeters such as pH, dosage, contact time, concentration, and agitation speed. The data from concentration and contact time are used to generate the equilibrium and kinetic data respectively. Data from equilibrium and kinetic studies are related to the CRH characterization studies to explain in detail the metal ion adsorption mechanism. Computational studies were done to further explain the interaction between the metal ions and the CRH with Schrodinger software, by carrying out docking studies using Glide. Schrodinger is a suite of molecular modeling packages that take advantage of the latest technological advances in computational chemistry . Additionally, pH studies were also carried out to account for the adsorption behaviors.
Equilibrium studies provide data on the capacity (or affinity) of the adsorbent for the solute and hence the dosage of absorbent required to remove a unit mass of solute from the solution. Mathematically equilibrium is expressed in isotherms. Two commonly used isotherms are Langmuir and Freundlich    and their applicability is judged with correlation coefficient (R2). The Langmuir isotherm assumes mono-layer adsorption on a uniform surface with a finite number of adsorption sites and maximum adsorption occurs when a saturated mono-layer of adsorbate is present on the adsorbent surface  . The linear form of Langmuir isotherm model  is given in Equation (1). Here, Kads is the equilibrium constant that increases with increasing molecular size (Langmuir constant), qmax is the adsorption capacity (mg/g), qe is the equilibrium adsorption (mg/g) and Ce is the equilibrium (residual) concentration of metal ion (mg/L).
The dimensionless separation factor RL is given in Equation (2) which is an essential feature of Langmuir isotherm . If RL = 0, the process is said to be irreversible, if RL = 1, it is linear, if RL > 1, it implies an unfavourable adsorption and if 0 < RL < 1, it’s a favourable adsorption.Ci is the initial concentration of metal ions (mg/L).
Freundlich isotherm is an empirical model and describes equilibrium on heterogeneous surface and does not assume mono-layer adsorption. It is an indicator of the extent of heterogeneity of adsorbent surface . The linear form of Freundlich isotherm  is represented in Equation (3), where K is Freundlich constant or adsorption capacity and 1/n is adsorption intensity.
Adsorption kinetics describes the adsorption mechanism as well as estimates the time required for obtaining equilibrium concentration of metal ions onto the biomass . It measures the efficiency of the process and shows a large dependence on physical and/or chemical characteristics of the sorbent material . Commonly used are pseudo-first-order and pseudo-second-order models and their linear forms are given in Equations (4) and (5) respectively .
Hereqt is the adsorption capacity at time t (mg/g) and k1 is the rate constant of pseudo-first-order (min−1) and k2 is the rate constant of pseudo-second-order (g·mg−1·min−1).
2. Materials and Methods
All the chemicals used in this study were analytical grade: CuSO4·5H2O (98%), Pb(NO3)2 (99%), H2SO4 (95%), NaOH (97%), (Ajax Fine Chem, Australia).
2.2. Preparation and Characterization of CRH of Bioadsorbent
Cassava samples were harvested from Markham, Morobe Province, PNG (−6˚38'29.51"S and 146˚51'37.55"E). The CRH were peeled from the pulp, cleaned and washed thoroughly with tap water and rinsed with distilled water. The CRH were then cut into smaller pieces, sun dried for 7 days and then ground and sieved to get particle sizes ≤ 50 µm. A summary schematic representation of the process is depicted in Figure 1.
CRH powder was characterized using FTIR (using JASCO FT/IR-4100 spectrometer, recorded as KBr pellet), CHN analysis (using PerkinElmer 2400 Series CHNS/O Analyzer), specific surface area and pore size volume (using Micromeritics ASAP 2020 surface area and porosimetry system) measured at Dept. of Chemistry, IIT Madras, India. The sample for specific surface area and pore size volume determinations was outgassed at 100˚C for 12 hour and the parameters were obtained using nitrogen adsorption BET isotherms. The equilibrium interval during the measurement was maintained for 5 sec. UV-Visible spectrum (using Varian 50 Bio UV-Vis spectrophotometer, recorded as a clear solution (1 g of powder was mechanically shaken with 100 mL of distilled water for 1 h and filtered)) and pH (using TPS Labchem pH meter at 28˚C ± 0.5˚C) were measured at Dept. of Applied Sciences, Papua New Guinea University of Technology (PNGUoT), PNG. Sulphur as sulphate, phosphorus as phosphate and metal ion concentrations were determined by ICP-OES using Varian 725-ES ICP-OES at UASL, Dept. of Agriculture, PNGUoT, PNG.
Figure 1. Preparation of CRH as bioadsorbent for the study.
2.3. Bioadsorption Experiment
Stock solutions (1000 mg/L) of metal ions were prepared by dissolving 3.92 g and 1.59 g of CuSO4·5H2O and Pb(NO3)2 respectively in distilled water. These solutions were diluted to various concentrations and were used in the experiments. The bioadsorption studies were carried out with the desired mass of the bioadsorbents placed in 100 mL of metal ion solution with desired concentration and pH in 250 mL conical flasks. The flask is then shaken with a mechanical flask shaker at 240 rpm at room temperature for the desired time. Buchner funnel with Whatman No.1 filter paper was used to filter the bioadsorbent with adsorbed metal and the filtrate was analyzed for residual metal ion concentration using inductively coupled plasma-optical emission spectroscopy (ICP-OES). The effect of the solution pH was investigated in the range of 1 - 12 with addition of 0.1/1M NaOH and H2SO4. Filtrates were collected at time ranging from 0.5 to 100 minutes to study the effect of the contact time. The effect of concentration was investigated with metal ion concentration ranging from 0.5 to 100 mg/L. Bioadsorbents dosage ranging from 0.05 to 10 g were used to investigate the effect of dosage.
3. Result and Discussion
3.1. Characterization of CRH
3.1.1. Infrared Spectral Studies
The FTIR spectra of CRH are provided in Figure 2. The presence of hydroxyl and carboxylic acid groups, methionine, and tryptophan, aromatic molecules such as thiamine and other amino acids   make the band assignments quite complex and an attempt was made to assign the observed bands (Table 1)
Figure 2. FTIR spectrum of CRH powder.
Table 1. Infrared spectral data for CRH and band assignments.
based on available literature . It is very clear that functional groups like acid (-COOH), amine (primary-NH2 and secondary-NHR), imine (>C=NH), alcohol (-OH), sulphydryl (-SH), mono-substituted and para-disubstituted benzene rings are available in CRH. Availability of these functional groups supports chemisorption process. .
Based on the infrared spectral studies, a better depiction of the ability of CRH to coordinate with metal ions is provided in Figure 3. Coordination could be considered as a heterogeneous phenomenon where CRH acts as a solid phase while the aqueous phase contains metal ions. Equilibrium was set up between the two species at the solid-solution interface. The abstraction of metals from aqueous solution is simply the coordination of metal ions to active ligating sites of CRH. This correlates with the results from pseudo-second-order kinetics studies indicating chemisorption as the driving force for adsorption.
3.1.2. UV-Visible Spectrophotometry
The oxalate ( , ox) present in CRH as a secondary metabolite  could possibly interact with M(II) [M = Cu(II) and Pb(II)] in aqueous solutions to give rise to a coordination complex, [M(ox)2]2− which are sparingly soluble in water. This was further proven by recording UV-Visible spectrum (Figure 4) of clear brown solution obtained by filtering the CRH suspension. Presence of peaks in the ultraviolet region between 190 - 366 nm set as an evidence for the presence
Figure 3. Possibilities of CRH functional groups to coordinate with metal ions [darkened portion: lone pairs of electrons for potential metal coordination].
Figure 4. UV-Visible spectrum of clear solution obtained from CRH suspension.
of oxalate in the solution. Oxalates are characterized by two prominent absorptions at 190 nm due to n → σ* and ~300 nm due to n → π* transitions . Oxalate is a polydentate ligand (bidentate) which gives rise to stable chelates upon sequestering M(II) . Because the complex was present in the filtrates which were subjected to ICP-OES analysis for metal estimations, some of the expected behaviors in the adsorption studies were not seen as reported by Jorgetto et al. .
3.1.3. Elemental Analysis
CHNSP analysis of CRH clearly supported the presence of sulphur, nitrogen and phosphorus containing groups which directly or indirectly aided in metal coordination. Table 2 provides a comparison of CHNSP data of the CRH from various studies. It can be concluded that the elemental compositions are variable, though they are all present in samples from various continents.
Table 2. CHNSP data for CRH powder from various studies.
3.1.4. Surface Area and Pore Volume
The BET surface area and the pore volume of CRH was found to be 0.9039 m2/g and 0.002101 cm3/g (based on single point adsorption on pores < 1019.662 Å radius) respectively. The result clearly corroborates with earlier findings and the surface area was found to be <1 m2/g . The data implies that the material is non-porous and hence rapid adsorption kinetics is expected which agreed well with the kinetic studies (Table 3).
3.2. Adsorption Studies
3.2.1. Experimental Parameters
The pH of the solution plays a very significant role in the adsorption process. For both Cu(II) and Pb(II), at lower pH, the adsorption of both metals was found to be low (Figure 4). This could be due to the diminished exchange of acidic sites of CRH with metals on the basis of Le Chatelier principle. At low pH, the hydrogen ion concentration is very high thus favoring the reverse reaction of Equation (6) and subsequently affecting the adsorption and removal processes of metal.
Beyond pH 7, the adsorption decreases, probably due to metals precipitating as hydroxides (Equation (7)). The maximum adsorption of Pb(II) and Cu(II) occurs at around pH 4 with removal of 95.26% and 78.66% respectively (Figure 4). However, the results are in contrast with the literature reported data  where the Cu(II) adsorption to cassava husk was found to be rapid from pH 1 - 3, beyond which the adsorption stabilizes to a plateau.
The effective dosage for Pb(II) was found to be 0.1 g of CRH in 100 mL of metal solution which adsorbed 95.2% (Figure 5). For Cu(II), the effective dosage was found to be 1 g in 100 mL of metal solution (78.98% adsorption). This indicates
Figure 5. Effect of pH on the adsorption of Pb(II) and Cu(II) to CRH.
Table 3. pH, specific surface are and pore volume data for CRH.
that the nature of metal ion is very vital for adsorption process because higher dosage (10 times more for Cu(II) compared to that of Pb(II)) as well as lower adsorption was observed for Cu(II).
The removal of Cu(II) and Pb(II) with CRH was fast (Figure 6). The adsorption equilibrium time for Pb(II) and Cu(II) was found to be 2 and 25 minutes respectively. The adsorption of both Pb(II) and Cu(II) onto CRH was high at lower concentration and decreases as the concentration increases (Figure 7). With 0.5 mg/L, Pb(II) and Cu(II) adsorption was found to be 94% and 76% respectively. Effects of pH, dosage, contact time and concentration on the adsorption of Cu(II) and Pb(II) ions to CRH are provided in Figures 5-8.
3.2.2. Equilibrium Studies
Table 4 below gives the concentration values (initial concentration, Ci, and equilibrium concentration, qe) for Cu(II) and Pb(II) solutions which were used to plot the Langmuir and Freundlich Isotherms.
1) Langmuir Isotherm
Table 4 gives the values of Ci and qe for the metals which were used to plot the isotherms. The Langmuir isotherm model for both metal ions is presented in Figure 9. The maximum adsorption capacities (qmax) for completely adsorbed metal monolayer coverage were found to be 26.74 and 16.50 mg/g of CRH for Pb(II) and Cu(II) ions respectively (Table 5). Kads is related to the affinity of the binding sites and lower values of Kads [0.0652 and 0.0175 for Pb(II) and Cu(II) respectively] indicated the particle radius for CRH was smaller towards adsorption
Figure 6. Effect of adsorbent dosage on the adsorption of Pb(II) and Cu(II) to CRH.
Figure 7. Effect of contact time on the adsorption of Pb(II) and Cu(II) to CRH.
Figure 8. Effect of concentration on the adsorption of Pb(II) and Cu(II) to CRH.
Figure 9. Langmuir isotherm for Pb(II) and Cu(II) adsorption to CRH.
Table 4. Equilibrium values for Pb(II) and Cu(II) adsorption.
Table 5. Langmuir isotherm and its coefficients.
. The separation factors (RL) for both Pb(II) and Cu(II) adsorption to CRH lie very close to 1 (Table 6) which supported its low R2 value for less favorable adsorption.
2) Freundlich Isotherm
The experimental data for Pb(II) and Cu(II) adsorption onto CRH showed a very good linear relationship with Freundlich isotherm model (Figure 10). This indicated that the adsorption reached equilibrium on a heterogeneous surface or the adsorption energies of CRH particles vary. Fitting the Freundlich isotherm also means the adsorption has multi layers of metals adsorbed onto CRH. As indicated in Table 7, the adsorption of Pb(II) and Cu(II) to CRH is favorable because it fulfills the condition of n > 1 . This data correlated the adsorption capacity, K value which was found to be 1.5485 and 0.2780 mg/g for Pb(II) and Cu(II) respectively. A very good linear relationship has indicated the strong binding of Pb(II) and Cu(II) ions to CRH .
3.2.3. Adsorption Kinetic Studies
1) Pseudo-First-Order Kinetics
The experimental data for the adsorption of Pb(II) and Cu(II) did not fit well
Table 6. Comparison of separation factors (RL) between Pb(II) and Cu(II) for adsorption with CRH.
Table 7. Freundlich isotherm and its coefficients.
Figure 10. Freundlich isotherm for Pb(II) and Cu(II) adsorption to CRH.
with Pseudo-first-order kinetic model (R2 = 0.2619 and 0.1993 respectively) as shown in Figure 11. No correlation was found for comparing the experimental equilibrium adsorption capacity (qe.exp) and calculated equilibrium adsorption capacity (qe.cal) (Table 8). It can be concluded that the adsorption process did not follow pseudo first order model .
2) Pseudo-Second-Order Kinetics
Figure 11. Pseudo-first-order kinetics for adsorption of Pb(II) and Cu(II) to CRH.
Table 8. Pseudo-first-order kinetics coefficients for Pb(II) and Cu(II) adsorption to CRH.
Both the adsorption of Pb(II) and Cu(II) to CRH has shown a very good linear relationship (R2 = 1 for both cases) with the pseudo-second-order kinetic model (Figure 12). Furthermore, qe.exp agreed well with qe.cal for Pb(II) but not so well with Cu(II) adsorption (Table 9). This agreed data implies that the adsorption depends on both the adsorbent and adsorbate and the rate limiting step was chemisorption process, involving valency forces through sharing or exchange of electrons between the adsorbent and adsorbate . The availability of functional groups on CRH (Figure 2) very well supports chemisorption of Pb(II) and Cu(II) through exchange or sharing of electrons.
3.3. Molecular Modeling Studies
A thorough literature search was undertaken to identify that there was no reported structure for CRH. Hence, data sets for 2600 unique protein sequences were screened to get a reasonable starting geometry for docking (coordination) studies. Molecular modelling studies were performed using Schrödinger software; a sophisticated chemical simulation software used in materials research . However, cassava protein sequences were reported  in an overview of protein identification studies. These sequences were used as the starting point for the modelling studies. Vanderschuren et al.  identified over 2600 unique proteins in the cassava root and provided data sets for these proteins. These data sets were screened to get the best possible match against the protein data bank to
Figure 12. Pseudo second order kinetics for Pb(II) and Cu(II) adsorption to CRH.
Table 9. Pseudo-second-order kinetic coefficients for adsorption of Pb(II) and Cu(II) to CRH.
get a reasonably good starting structure for our docking studies with metal ions. Docking studies using Glide predicted several sites in the structure capable of docking with metal ions (Figure 13).
3.4. Coordination Ability of CRH to Metals
The properties of metals that influence the coordination possibilities with the ligating atoms and hence the stability of coordination complex formed are 1) Nature; 2) Size, charge and charge density; 3) Electronic configuration; and 4) Hardness . The properties of ligating atoms that decide the ease with which they coordinate with the metal are 1) Lonepair availability on the ligating atoms (N, O, S, etc.) and hence the basicity; and 2) Hardness of ligating atoms . According to Pearson’s hard and soft acids and bases (HSAB) concept , hard metals (hard Lewis acids) form stable complexes with hard ligating atoms (hard Lewis bases) and vice versa. According to HSAB, both Cu(II) and Pb(II) ions could be classified as intermediate or borderline acids; similarly, CN−, >C=O, S2− and RS− are considered as soft bases while O2−, –NH2, and –C(O)O− are considered as hard bases . However, considering the ionic radii of the two metals [Pb(II) 119 pm versus Cu(II) 73 pm], Cu(II) is smaller in size and hence easily polarizes the larger ligands compared to Pb(II). Thus Pb(II) could be relatively a softer Lewis acid compared to Cu(II), though both were originally grouped
Figure 13. Structure of a model of cassava root husk docking with metal ions.
under borderline category. This electron polarization can possibly induce covalency on the bond formed, thus imparting stability to the species (coordinated complex) formed. The presence of varying ligating sites in CRH ensures that their coordination with both Cu(II) and Pb(II) occurs to differing extent based on aforesaid properties of ligands and metals. These facts were also supported by the molecular modeling studies.
The adsorption study was carried out using CRH powder by varying the parameters such as pH, adsorbent dose, contact time, and concentration of metals. The study brought out the fact that this bioadsorbent was very effective in removing Pb(II) from aqueous solutions compared to Cu(II) ions and this was attributed to the soft nature of the former metal and the very high charge density of the latter. This research clearly has concluded that the metal removal processes by adsorption are very much dependent on the nature, surface charge density, and hardness of the metals, and the nature of sites available on the bioadsorbents for binding. This study has conducted trials pertaining to metal removal from aqueous solutions using naturally available bioadsorbents and correlates the results with the coordination ability of metals and the availability of ligating sites on bioadsorbents chosen. Attempts were made to assign FTIR spectral bands which also indirectly provide evidences for the presence of ubiquitous amounts and types of amino acids in the chosen bioadsorbent.
Conceptualization, JG and KP; methodology, KP; formal analysis, KP.; investigation, KP, JG and RJ; resources, JG; data curation, KP and JG; writing—original draft preparation, KP; writing—review and editing, JG; visualization, JG, KP and RJ; supervision, JG; project administration, JG; funding acquisition, JG. All authors have read and agreed to the published version of this article.
This research was funded by Papua New Guinea University of Technology Research Funds.
KP thanks the Research Committee, PNGUoT for scholarship for pursuing M. Phil. Degree, Dr. Sivakumar Balakrishnan for guidance and generosity in editing this paper and Dr. Yalambing for approving the article processing charge. JG thanks the Research Committee, PNGUoT for the project funds. Dept. of Applied Sciences and Dept. of Agriculture, PNGUoT and SAIF, IITM are gratefully acknowledged for spectral data collected.
 Papua New Guinea Constitutional Law Reform Commission (2013) Review of Environmental and Mining Laws Relating to Management and Disposal of Tailings.
 Khraisheh, M., Al-Ghouti, M.A. and AlMomani, F. (2020) P. putida as Biosorbent for the Remediation of Cobalt and Phenol from Industrial Waste Wastewaters. Environmental Technology & Innovation, 20, Article ID: 101148.
 Sulyman, M., Namiesnik, J. and Gierak, A. (2017) Low-Cost Adsorbents Derived from Agricultural By-Products/Wastes for Enhancing Contaminant Uptakes from Wastewater: A Review. Polish Journal of Environmental Studies, 26, 479-510.
 Egbosiuba, T.C., Abdulkareem, A.S., Roos, W.D., Afolabi, E.A., Tijani, J.O., Bankole, M.T., et al. (2021) Adsorption of Cr(VI), Ni(II), Fe(II) and Cd(II) Ions by KIAgNPs Decorated MWCNTs in a Batch and Fixed Bed Process. Scientific Reports, 11, Article No. 75.
 Ibrahim, B. (2021) Heavy Metal Ions Removal from Wastewater Using Various Low-Cost Agricultural Wastes as Adsorbents: A Survey. Zanco Journal of Pure and Applied Sciences, 33, 76-91.
 Amar, M.B., Walha, K. and Salvado, V. (2021) Valorisation of Pine Cone as an Efficient Bioadsorbent for the Removal of Pb(II), Cd(II), Cu(II), and Cr(VI). Absorption Science and Technology, 2021, Article ID: 6678530.
 Information Technology Service (n.d.) Scientific Application-Schrodinger. The University of North Carolina, Chapel Hill.
 Alene, A.N., Abate, G.Y., Habte, A.T. and Getahun, D.M. (2021) Utilization of a Novel Low-Cost Gibto (Lupinus albus) Seed Peel Waste for the Removal of Malachite Green Dye: Equilibrium, Kinetic, and Thermodynamic Studies. Journal of Chemistry, 2021, Article ID: 6618510.
 Banerjee, K., Ramesh, S.T., Gandhimathi, R., Nidheesh, P.V. and Bharathim, K.S. (2012) A Novel Agricultural Waste Adsorbent, Watermelon Shell for the Removal of Copper from Aqueous Solutions. Iranica Journal of Energy & Environment, 3, 143-156.
 Hossain, M.A., Ngo, H.H., Guo, W.S. and Nguyen, T.V. (2012) Biosorption of Cu(II) from Water by Banana Peel Based Biosorbent: Experiments and Models of Adsorption and Desorption. Journal of Water Sustainability, 2, 87-104.
 Berrazoum, A., Marouf, R., Ouadjenia, F. and Schott, J. (2015) Bioadsorption of a Reactive Dye from Aqueous Solution by Municipal Solid Waste. Biotechnology Reports, 7, 44-50.
 Ezeonuegbua, B.A., Machidoa, D.A., Whonga, C.M.Z., Japhet, W.S. Alexiouc, A., Elazab, S.T., Qusty, N., Yaro, C.A. and Batihag, G.E.S. (2021) Agricultural Waste of Sugarcane Bagasse as Efficient Adsorbent for Lead and Nickel Removal from Untreated Wastewater: Biosorption, Equilibrium Isotherms, Kinetics and Desorption Studies. Biotechnology Reports, 30, Article No. e00614.
 Ngah, W.S. and Hanafiah, M.A.K.M. (2008) Removal of Heavy Metal Ions from Wastewater by Chemically Modified Plant Wastes as Adsorbents: A Review. Bioresource Technology, 99, 3935-3948.
 Jorgetto, A.O., Silva, R.I.V., Saeki, M.J., Barbosa, R.C., Martines, M.A.U., Jorge, S.M.A., Silva, A.C.P., Schneidr, J.F. and Castro, G.R. (2014) Cassava Root Husks Powder as Green Adsorbent for the Removal of Cu(II) from Natural River Water. Applied Surface Science, 288, 356-362.
 Lukuyu, B., Okike, I., Duncan, A., Beveridge, M. and Blummel, M. (2014) Use of Cassava in Livestock and Aquaculture Feeding Programs. ILRI Discussion Paper 25, International Livestock Research Institute, Nairobi, 1-83.
 Batista de Souza, C.R., Dos Reis, S.P. and Carvalho, L.J.C.B. (2015) An Overview of Protein Identification Studies in Cassava. Current Protein and Peptide Science, 16, 219-227.
 Vanderschuren, H., Nyaboga, E., Poon, J.S., Baerenfaller, K., Grossmann, J., Hirsch-Hoffmann, M., Kirchgessner, N., Nanni, P. and Gruisse, W. (2014) Large-Scale Proteomics of the Cassava Storage Root and Identification of a Target Gene to Reduce Postharvest Deterioration. Plant Cell, 26, 1913-1924.