Computational exploration of bio-remediation solution for mixed plastic waste

The plastic materials are recalcitrant in the open environment, surviving longer without complete remediation. The current disposal methods of used plastic material are not efficient; consequently, plastic wastes are infiltrating the natural resources of the biosphere. A sustaining solution for plastic waste is either recycling or making it part of the earth’s biogeochemical cycle. We have collected, manually mined, and analyzed the previous reports on plastic biodegradation. Our results demonstrate that the biodegradation pattern of plastics follows the chemical classification of plastic types. Based on clustering analysis, the distant plastic types are grouped into two broad categories of plastic types, C-C (non-hydrolyzable) and C-X (hydrolyzable). The genus enrichment analysis suggests that Pseudomonas and Bacillus from bacteria and Aspergillus and Penicillium from fungal are potential genera for bioremediation of mixed plastic waste. Overall results have pointed towards a possible solution of mixed plastic waste either in a circular economy or open remediation. The meta-analysis of the reports revealed a historical inclination of biodegradation studies towards C-X type of plastic; however, the C-C class is dominated in overall plastic production. An interactive web portal of reports is hosted at plasticbiodegradation.com for easy access by other researchers for future studies


38
Plastic is a collective term used for multiple polymeric materials known to be 39 lightweight, durable, and simultaneously can own bespoke shapes. A polymeric material is an 40 extended repetition of the characterized monomeric unit synthesis from nature or chemical 41 origins. Besides having advanced materialistic properties, they also provide an economical and 42 real-time supply edge over alternatives such as glass, steel, and bio-based fibers. Naphtha and 43 natural gas are the leading suppliers of plastic monomers, and it accounts for 6% of total annual 44 crude oil productions (Geyer et al., 2017). There are around 50 main plastic types; however, 45 their annual production is not uniform but skewed towards a few popular plastic types ( underutilized (Vollmer et al., 2020). The typical approach of reducing, redesigning, reusing, 71 and recycling is not working because this is not making any dent in the futuristic production of These recycling efforts do not resupply the same plastic types used in the process but instead 75 become the source of downgraded plastics; meanwhile, the supply of recycled plastic types is 76 continuously coming from virgin plastic synthesis. The single-use plastic contributes around 77 50% of the overall plastic production and frequently ends up in the open environment due to 78 poor incentives to recycle and irresponsible consumer behaviour. It has become the most-79 talked-about category of plastic in terms of regulation and management (Xanthos & Walker,80 2017; Elliott et al., 2020). The current global supply chain has made the scenario even more 81 complex as the producer and end-user often fall in different areas of the globe, so any legal 82 options to pass the liability on producer has a thin margin for emerging ( interactive portal on plasticbiodegradation.com for the appropriate users to explore. The 172 interactive interface of our database will be helpful for anyone to start literature mining on any 173 aspect of microbial degradation of plastic materials. 174 175

177
Microbial degradation result 178 179 The microbial degradation of plastic wastes in an open environment follows the process 180 of biodeterioration, depolymerization, and assimilation. Biodeterioration of plastic material is 181 the initial process to dismantle the polymeric surface through extracellular polymeric 182 substances to provide the optimum condition for anchoring microorganisms. Microbes use the 183 extracellular release of internal metabolites to change the physicochemical condition, such as 184 surface pH, to obtain deterioration (Sharma et al., 2017). The non-biotic forces of the 185 environment, such as sunlight, water, and air, also assist in priming the initial conditions. Once 186 the microbe establishes the biofilm on the plastic surface, they release extracellular enzymes 187 to depolymerize the plastic polymer. The depolymerization process converts more extended reports have both bacteria and fungi species which were also included in study. The ubiquitous 196 presence of bacterial degradation studies could be due to robust survival capacity of bacterial 197 species and easy experimental handling of bacterial based biodegradation.

199
We have analyzed the taxonomical data of microbes at the Genus level to circumvent the 200 differences in various studies at their taxon level identification. In our in-house data we have 201 218 unique genera for 80 different plastic types, and only 62 genera are reported for more than 202 three plastic types. We only considered those plastic types reported to be degraded by more 203 than ten unique genera, which resulted in 20 plastic types which we considered for our analysis 204 out of 80 different types of plastic. The PLA/PHB blend was dropped from further analysis due 205 to mixed composition. An association matrix of size (207x19) was filled with the binary values 206 of 1 and 0 based on genus plastic types occurrence in metadata. The association matrix is 207 further utilized to get a co-occurrence matrix of size (19x19) for all plastic-type. The Figure 2  208 shows the non-metric multidimensional scaling (NMMDS) and hierarchical clustering plots of 209 selected plastic types with precomputed cosine similarity of co-occurrence matrix. The 210 selected plastic types are clustered based on their membership of C-C or C-X type of plastic 211 substrates. The results presented in Figure 2 indicate that PP, LDPE, HDPE, PE, and PS falls 212 in one cluster which appears to be specific to C-C type of plastics, whereas PHB, PHBV, PLA, 213 PES, PET, PCL, PBS, PBSA are forming another cluster of C-X type plastic material. The 214 plastic type PU, PHA, and PHBH didn't fall in any cluster, therefore, behave as outliers, while 215 Nylon appears to be closer to C-C type of plastic in their biodegradation behaviour.

217
To decode the preference of bacterial genus for plastic types, we did enrichment 218 analysis. For the same, we have selected 48 genera of bacterial origin which were reported for 219 more than two types of above-mentioned selected plastics. The results are presented in Figure  220 3 as a heatmap of bacterial genus and plastic types. The colour intensity in the heatmap 221 normalized on total entries of the column for individual plastic-type. The results indicate that 222 the microbial degradation of plastic is not concentrated to a few specific genera but spread all 223 over in the plot. The careful observation of the heatmap demonstrates that genus of 224 Betaproteobacteria class (Roseateles to Achromobacter) are enriched for C-X class of plastic 225 types, specifically PHAs. Alphaproteobacteria (Brevundimonas and Rhodopseudomonas) and 226 Gammaproteobacteria class (Alcanivorax to shewanella) are enriched for C-C class of plastic 227 material except Psuedomonas which appears to be selective for both C-C and C-X mimicking 228 Bacilli class; the genus of Bacilli class (Laceyella to Staphylococcus) are active for both the C-229 C and C-X type plastic material. Overall, the genus of Actinomycetia class (Thermobifida to 230 Micrococcus) is less enriched, however, appears to be selective for the C-X type plastic 231 materials expect Rhodococcus. and The Streptomyces appears to be selective for multiple C-232 X type of plastic materials.

234
Further, to decode preferential selectivity of various fungal genus for diverse sets of 235 plastic material, we performed enrichment analysis. Of all 79 genera of fungi kingdom 236 associated with plastic degradation; only 28 are reported for more than two plastic-type. Figure  237 S1 shows the heatmap of fungal genus and plastic types. We selected 42 genera of fungal origin 238 reported for more than one selected plastic-type. Although, PHBH has no degradation report 239 from the fungi genera therefore it is not mentioned in Figure S1. Overall the available 240 microbial biodegradation data indicates that a smaller number of fungal genus are reported 241 which can degrade plastics compared to bacterial genus. The data presented in FigureS1, 242 indicates that genus of the class Eurotiomycetes has been extensively reported for plastic 243 degradation; among them, the genus Aspergillus and Penicillium appears to degrade a variety 244 of plastic materials across C-C and C-X category. According to heatmap presented in Figure  245 S1, Phanerodontia genus appears to be enriched for the unsaturated C-C type plastics. The 246 Fusarium genus also shows degradation capability for multiple plastic types, specifically the 247 aliphatic ester group containing C-X class of plastic materials.  are reported against only a single plastic-type. Figure S2 shows the number of enzymes for 266 different plastic types in descending order and the distribution is observed to be skewed 267 towards C-X type of plastics. The results indicate that highest numbers of enzymes are reported 268 for PCL, PET, and PHB type plastic.

270
The To enrich the sparse enzyme-plastic association matrix, we used the sequence similarity 285 linkage to assign a plastic-type to a new enzyme with a higher E-value for two or more already 286 known enzymes for that plastic type. The local all vs. all blast method has been performed on 287 153 enzymes, and it added 50 new connections between enzymes and plastic types at the 288 sequence similarity threshold of e value e -150 . Figure 4 shows the NMMDS and hierarchal 289 clustering plots of enzyme-plastic association matrix with precomputed cosine similarity. The 290 overall pattern of enzymatic degradation of plastic types reiterates the earlier classification and 291 results appears to segregate the enzymes into two bigger groups specific for C-C and C-X 292 backbone classes. The hierarchical clustering and multi-dimentional scattering data presented 293 in Figure 4 demonstrate that C-X class is broadly divided into two sub-groups, one dominated 294 by PHAs and the other has the rest of C-X type of plastics. Enzyme degrading PHAs also 295 degrade PES showing the similarities in catalytic property, therefore, PES in an outgroup 296 member in PHAs dominated group. The PU and Nylon both follow their outgroup behaviour 297 and come close to the C-C class of plastic types which is as per our microbial degradation data 298 reported in Figure 2. The PEG is within the C-X group of plastic types; however, it shows the 299 distance from other polyester backbone types. 300 301 Co-occurrence result 302 The co-occurrence analysis of the group of entities is a standard technique to deduce the 303 relationship between them in the field of natural language processing (NLP) (Rozmus, 2009). 304 We perform the co-occurrence analysis on our in-house plastic biodegradation data to find the 305 pattern of similarity between different plastic types. There are 481 reports for 81 different 306 plastic types; out of which 71% of studies report biodegradation of single plastic type. The 307 Figures S3 and S4 show frequency distribution of the plastic types reported in previous studies. 308 For co-occurrence analysis, we could select 21different plastic-type which were reported for 309 more than five studies. A publication and plastic-type association matrix of size (446x21) is 310 filled with binary entries based on their occurrence in the in-house database record. Later, the 311 association matrix is used to calculate a cosine similarity matrix of size (21x21) for all selected 312 plastic types. Figure 5 shows the NMMDS and hierarchal clustering plots of co-occurrence of 313 different plastic types with precomputed cosine similarity. The data clearly indicates two bigger 314 clusters corresponding to two main plastic-type category (C-C and C-X type plastics). The C-315 C group of plastic includes PE, PS, PP, LLDPE, HDPE, and LDPE; however, PVC and PVA 316 show outgroup behaviour. The C-X group include PHO, PHB, PHBV, PLA, PBS, PBSA, PCL, 317 and PES plastic types while nylon, PU, and PEG behave as outgroups compared to the rest of 318 the C-X members. Although PET and PBAT being aromatic polyester display closeness, their 319 similarity is not substantial. The overall co-occurrence pattern mimics the microbial and 320 enzymatic degradation analysis patterns and supports the earlier classification of plastic types 321 (C-X and C-C type of plastic). 322 323

324
Plastic materials have been proved to be an outstanding discovery of modern time; (https://www.mckinsey.com/industries/chemicals/our-insights/how-plastics-waste-recycling-334 could-transform-the-chemical-industry). 335 Recycling used plastic is a potential alternative to virgin plastic; however, producers 336 are not interested in recycling every plastic type due to functional and economic factors ( hydrolyzable(C-X) and non-hydrolyzable (C-C) groups. The results of all analysis shows that 353 the pattern of microbial degradation follows the class segregation of plastic-type. Plastics with 354 hydrolyzable and nonhydrolyzable backbones behave differently but are similar within the 355 group with few exceptions such as PU, Nylon, and PET. Figure S5 shows the annual frequency 356 of reports based on classification into hydrolyzable(C-X) and non-hydrolyzable (C-C) groups.

357
It reveals a historic inclination towards the C-X (hydrolyzable) group of plastic-type. The pre-358 treatment of plastic material with physical or chemical methods can influence precise microbial 359 degradation behaviour and, therefore, weaken the clustering patterns and lower the statistical 360 power to determine the clusters in data. Meanwhile, lack of proper treatment reporting in all 361 publications is the reason that we did not consider the pre-treatment factor in our analysis.

363
We analysed the taxonomical data of microbial strains at the genus level due to 364 incomplete information for species level, although strains from the same Genus can vary on 365 genomic content based on their collections site. Genus Aspergillus, Pseudomonas, and Bacillus 366 would be perfect candidates to explore a solution for mixed plastic waste. Rhodococcus genus 367 is enriched only for nonhydrolyzable backbone plastic types that would be an excellent 368 contender for further exploration as they have a majority of stakes in overall production. The 369 higher diversity at the genus level restricts the interpretability of our findings. This ambiguity 370 at the genus level can only indicate the potential solutions, but search space is not limited to 371 those genera. There are more enzymatic degradation reports for hydrolyzable plastic types than 372 nonhydrolyzable C-C backbone types as later requires action from a cocktail of enzymes for 373 degradation. The enzymatic degradation patterns of PHAs are different from other C-X 374 backbone plastic types. The overall analysis showed that whole cell-based treatments are more 375 successful than enzyme-based treatments for nonhydrolyzable C-C backbone types plastic.

377
The results of co-occurrence analysis are similar to microbial and enzymatic degradation 378 analysis. The PEG, PBAT, PET, Nylon, PU, and PVA show outgroup behaviour identical to 379 previous results. It appears that these plastic types represent an intermediatory group in 380 between both C-C and C-X groups based on their biodegradation behaviour. Most of the 381 reports are about the single plastic type, which led to a sparse co-occurrence matrix. Future 382 studies with multiple plastic types are strongly recommended to have higher possibilities to 383 find a generic solution for mixed plastic waste. Success in finding a microbial strain that is 384 simultaneously effective on the whole plastic class, either C-C or C-X, would be taken further 385 to a biotechnology-based solution of mixed plastic waste. Nevertheless, a consortium of 386 microbial strains is a more feasible solution for an open environment degradation of the mixed 387 plastic-type. The bioremediation based solutions are successfully applied in the case of a 388 marine oil spill and heavy metal contamination (Abatenh et al., 2017). A similar solution would 389 be suitable for high concentration zones of plastic waste, such as landfills and metropolis 390 drainages. A practical and inexpensive solution to plastic waste is required for a sustainable 391 future. 392 393

Conclusion 394
Plastic waste has long-lasting tolerance against nature's physicochemical and biological 395 forces. The unbridled production of plastic materials combined with the irresponsible 396 behaviour of the end-user has accelerated the harmful impacts of plastic materials in the 397 ecosystem. A collective effort is required at different aspects of plastic material from 398 composition to disposal to prevent this catastrophe. The scientific community has been 399 searching for a bio-based solution for plastic waste. There are multiple reports of plastic 400 microbial degradation, specifically from highly contaminated regions such as landfills. Based 401 on these reports, the big corporations promote some types of plastics types as biodegradable to 402 influence consumer's behaviour; however, this is not entirely true. Large-scale eco-friendly, 403 efficient methods have not yet been discovered to tackle the huge accumulation of used plastic 404 waste. In this research work, We have done a meta-analysis on the reports of microbial 405 degradation of various plastics. The cluster analysis of microbial degradation, enzymatic 406 degradation, and co-occurrence of various plastic types strengthens plastic's earlier 407 classification into hydrolyzable(C-X) and non-hydrolyzable (C-C) groups. We observed that 408 the number of reports from bacterial degradation is more than the fungal, and within the 409 kingdom, the distribution is more concentrated on particular genera. The metadata genus 410 enrichment analysis points towards a few putative genera for biodegradation solutions for 411 mixed plastic waste. The meta-analysis indicated that a whole cell-based solution is more 412 suitable for C-C groups as their enzymatic reports are significantly less than the C-X group of 413 plastic types. Most of the reports contain fewer than five plastic types, which hinders the 414 chances of finding a candidate through any metaanalysis which are capable of degrading 415 multiple types simultaneously. The molecular mining of such strains will give insight into 416 further engineering to develop large-scale solutions for mixed plastic. There are large 417 discrepancies in reporting the identity of microbial strains, plastic pretreatment, and overall 418 protocol in quantifying biodegradation in these studies. The scientific community needs to 419 develop a standard protocol for reporting plastic biodegradation findings. An established 420 protocol would assist in comparing and utilizing previous findings for future studies. We have 421 compiled all the reports into an interactive web portal at www.plasticbiodegradation.com for 422 easy access, which will be regularly updated in the coming future.

561
Characterizing the multidimensionality of microplastics across environmental compartments.