Characterizing genomic variants and mutations in SARS-CoV-2 proteins from Indian isolates

Abstract

SARS-CoV-2 is mutating and creating divergent variants by altering the composition of essential constituent proteins. Pharmacologically, it is crucial to understand the diverse mechanism of mutations for stable vaccine or anti-viral drug design. Our current study concentrates on all the constituent proteins of 469 SARS-CoV-2 genome samples, derived from Indian patients. However, the study may easily be extended to the samples across the globe.

We perform clustering analysis towards identifying unique variants in each of the SARS-CoV-2 proteins. A total of 536 mutated positions within the coding regions of SARS-CoV-2 proteins are detected among the identified variants from Indian isolates. We quantify mutations by focusing on the unique variants of each SARS-CoV-2 protein. We report the average number of mutation per variant, percentage of mutated positions, synonymous and non-synonymous mutations, mutations occurring in three codon positions and so on. Our study reveals the most susceptible six (06) proteins, which are ORF1ab, Spike (S), Nucleocapsid (N), ORF3a, ORF7a, and ORF8. Several non-synonymous substitutions are observed to be unique in different SARS-CoV-2 proteins. A total of 57 possible deleterious amino acid substitutions are predicted, which may impact on the protein functions. Several mutations show a large decrease in protein stability and are observed in putative functional domains of the proteins that might have some role in disease pathogenesis. We observe a good number of physicochemical property change during above deleterious substitutions.

1. Introduction

Due to the massive outbreak of COVID-19 disease, caused by the highly infectious novel coronavirus- SARS-CoV-2, the world is passing through a difficult situation. There are seven species of human coronaviruses reported so far that causing diseases in humans. Out of them, four species (HCoV-229E, HKU1, NL63 and OC43) causing mild respiratory apparatus infection which can easily be treated. However, three species, termed as beta coronaviruses (SARS-CoV, MERS-CoV, and SARS-CoV-2), are severe in nature, leads to potentially fatal consequences (Andersen et al., 2020). The scientific community trying hard to decipher parthenogenesis mechanism of SARS-CoV-2 and its therapeutic control, in silico, using various computational tools. An exhaustive study is available in (Das et al., 2020a).

Scientists observed a number of variants among novel coronavirus, SARS-CoV-2, reported from different geographical regions (Joshi and Paul, 2020; Sardar et al., 2020; Chang et al., 2020). Most of the evolutionary changes in the genome of viruses occur due to mutation. In some cases, it is due to insertion or deletion in the genome. In the course of evolution, variations bring novelty (Baer, 2008). The small variations might be beneficial or detrimental for the organism (Loewe and Hill, 2010). The mutational study helps in understanding viral transmission, replication efficiency, and magnitude of virulence of the pathogen (Eaaswarkhanth et al., 2020). A minor change in the genome might lead to the variation in functionality of constituent proteins of the organism (Chaudhuri, 2020). Previous studies revealed significant alternation in structural and pathogenic properties due to even single point mutation in virus proteins (André et al., 2019; Sakai et al., 2017). Characterizing mutations in different functional domains of SARS-CoV-2 genome might help in designing potential vaccine (Kaur et al., 2020).

Determining the mutation types (synonymous or non-synonymous) that influence a lot in gene regulation is vital for understanding the role of regulatory variation during evolution (DiMaio and Nathans, 1982; Foy et al., 2003). Studying the mutations at different codon positions is essential, particularly for quantification of synonymous and non-synonymous amino acid substitutions (Plotkin and Kudla, 2011). Though the non-synonymous mutation is primarily crucial (from codon usage bias point of view) as it alters the amino acid, synonymous mutations too have their strong impact (Plotkin and Kudla, 2011; Kristofich et al., 2018; Gustafsson et al., 2004). It is worth to mention that the changes in the physicochemical properties of nucleotides (purine-R or pyrimidine-Y) due to the mutations have remarkable biological significance (Lyons and Lauring, 2017; Sengupta et al., 2018; Guo et al., 2017). It is reported that in the case of codons, various evolutionary constraints at different codon positions occur due to the functional constraints imposed by the genetic code and the physicochemical properties of encoded amino acids (Bofkin and Goldman, 2007; Simmons, 2017; Plotkin and Kudla, 2011). For example, mutations at the 2nd position of a codon directly impact the changes in replaced amino acids (hydrophobic to hydrophilic and vice versa). The change is due to the transversion (A ↔ C or A ↔ T or G ↔ C or G ↔ T) (Haig and Hurst, 1991; Wolfenden et al., 1979; Błażej et al., 2017), although A ↔ G or C ↔ T transition is mostly occurring for single point mutation (Beletskii and Bhagwat, 1996; Błażej et al., 2017). Further, the changes in physicochemical properties of amino acids have a significant functional role (Das et al., 2019; Basak et al., 2017). Hence, understanding the genetic diversity is important that might hint towards the susceptible antigen targets of SARS-CoV-2. It can be used for potential therapeutic and prophylactic interventions in order to prevent this deadly outbreak. Mutations in SARS-CoV-2 proteins may lead to different phenotypic changes, and hence virus can adapt to new hosts and environments. In addition, codon bias study helps in revealing the host-virus interaction mechanism in SARS-CoV-2 (Dilucca et al., 2020; Kurland, 1991; Das et al., 2020b).

A detailed in-silico study on putative mutations in SARS-CoV-2 is of utmost important to understand any significant pattern and its possible impact on the functional and structural characteristics of the virus.

India is the second-largest SARS-CoV-2 infected country in the world. Due to the volume of study, we restricted our current study within Indian isolates only. However, our study can easily be extended to other variants from any part of the world. Although a couple of studies have been carried out to learn various crucial facts about SARS-CoV-2 genome from Indian patient samples (Kaur et al., 2020; Saha et al., 2020; Samaddar et al., 2020), there are certain facts yet to explore. Therefore, in this work, we broadly focused on the mutational study on SARS-CoV-2 genomes, isolated from Indian patient, as discussed in the following section.

2. Material and methods

2.1. Collection of SARS-CoV-2 genome sequences extracted from Indian patients

We collect SARS-CoV-2 genome sequences isolated from Indian patients that are achieved in public repositories. Several protein-coding genes are present in each SARS-CoV-2 genome. SARS-CoV-2 encodes different types of essential proteins: (i) nonstructural proteins - polyprotein (ORF1ab), structural proteins - Spike glycoprotein (S), Envelope (E), Membrane (M) and Nucleocapsid (N), and accessory proteins - ORF3a, ORF6, ORF7a, ORF7b, ORF8, and ORF10 (Kim et al., 2020; Yadav et al., 2020; Ruan et al., 2003; Gordon et al., 2020). A complete topological structure (position) of all SARS-CoV-2 proteins is shown in Table 1. Each of these proteins is highly essential and has diverse functional roles. The first full genome sequence of SARS-CoV-2 virus from India sample was reported during February 2020 (Yadav et al., 2020). We collect sequences from NCBI database1 (Supplementary-1). We find around 469 complete SARS-CoV-2 nucleotide sequences. Protein wise, we extract the coding region from each nucleotide sequence and ignore noisy sequences. The final list of obtained unique sequences is utilized for sub-sequence analysis (Table 2).

Table 1.

Topological structure of all SARS-CoV-2 proteins shown by respective genomic location. For each protein, the range of CDS region and amino acids used in this paper are numbered starting from 1 to length of the nucleotide or protein sequence.

Gene/protein	Genome location (nucleotide)	Protein length (aa)	Nucleotide location used	Amino acid location used
ORF1ab	266–21,555	7096	1–21,288	1–7096
S	21,563–25,384	1273	1–3819	1–1273
ORF3a	25,393–26,220	275	1–825	1–275
E	26,245–26,472	75	1–225	1–75
M	26,523–27,191	222	1–666	1–222
ORF6	27,202–27,387	61	1–183	1–61
ORF7a	27,394–27,759	121	1–363	1–121
ORF7b	27,756–27,887	43	1–129	1–43
ORF8	27,894–28,259	121	1–363	1–121
N	28,274–29,533	419	1–1257	1–419
ORF10	29,558–29,674	38	1–114	1–38

Table 2.

The number of collected samples from Indian isolates, unique variant, sample to variant ratio in each SARS-CoV-2 protein.

Protein	# collected samples	# noise free samples	# unique variant	Sample to variant ratio
ORF1ab	462	400	262	1.52
E	460	445	3	148.33
M	460	457	18	25.38
N	463	455	53	8.58
S	462	436	90	4.84
ORF3a	459	445	33	13.48
ORF6	460	459	3	153.0
ORF7a	460	454	11	41.27
ORF7b	456	455	3	151.66
ORF8	461	451	11	41.00
ORF10	460	460	3	153.33

2.2. Workflow design

Protein specific nucleotide sequences are first clustered to extract set of unique sequences (or unique variants). Next, unique sequences (representative of each group) are aligned using multiple sequence alignment. As a reference sequence we use sequence of SARS-CoV-2 proteins from Wuhan-Hu-1 (accession no: NC_045512). We compare every variant with the reference sequence to identify and localize mutations. We consider only single point mutation as a substitution. Observed mutations are then analyzed based on the number of synonymous and non-synonymous substitutions, quantification of nucleotide mutations in three different codon positions (1st/2nd/3rd), type of nucleotide mutations and amino acid substitutions. We then characterize non-synonymous amino acid substitutions and their biological implications using various computational tools.

2.3. Computational tools and techniques used

We use web-based tool PROVEAN2 and I-mutant3 for functional assessment of single point mutation. PROVEAN (Protein Variation Effect Analyzer), a web server, is used to predict any non-synonymous amino acid substitution or indel impacts on the biological function of a protein (Choi et al., 2012). The tool predicts two kinds of substitution effects: deleterious effect and neutral effect on protein function by measuring the combined score of substitution matrix, alignment, the position of substitution with the neighborhood that surrounds the site of variation. The cut-off value of the PROVEAN score is set as −2.5, below which it indicates deleterious substitution, otherwise, neutral. For predicting stability changes due to mutation, we use I-Mutant (Capriotti et al., 2005). The tool is designed based on Support Vector Machine (SVM) that produces Gibbs free energy of unfolding (ΔΔG value in kcal/mol, in terms of increased or decreased stability) for each non-synonymous substitution. The stability predictors value ΔΔG < − 0.5 indicates high decrease in stability, whereas, ΔΔG > 0.5 indicates high increase in stability, and −0.5 < ΔΔG ≤ 0.5 signifies neutral stability.

We use simple Python scripting for rest of the quantitative analysis.

We report the functionally important mutations identified using the above tools, highlighting the various putative functional domains of SARS-CoV-2 proteins. We also study wild type and new amino acid changes in two categories of physicochemical properties, Hydropathy profile (Aftabuddin and Kundu, 2007), and side-chain structure (Das et al., 2016). The categorizations are as follows:

• •Hydropathy based classes: The three classes are Hydrophobic (F, M, W, I, V, L, P, A), Hydrophilic (N, C, Q, G, S, T, Y), and Charged (R, D, E, H, K).
• •Side-Chain based classes: According to this grouping, twenty (20) amino acids are clustered into eight groups as Acidic (D, E), Basic (R, H, K), Aromatic (F, W, Y), Aliphatic (A, G, I, L, V), Cyclic (P), Sulfur-containing (C, M), Hydroxyl-containing (S, T), and Acidic amide (N, Q).

3. Results and discussion

Our first objective is to find out unique variants by clustering the SARS-CoV-2 gene sequences. We then identify point mutation (as substitution) in each observed variant by comparing it with the reference sequence. Observed mutations occurring at different codon positions are then classified and quantified based on different perspectives as discussed below.

3.1. Clustering of unique variants

The majority of the input genomes are redundant with respect to sequence similarity. We cluster them based on sequence similarity and consider a sequence from each cluster as cluster representative (termed as unique variant). We use a string matching technique to cluster the sequences, where exactly similar sequences are put in a single cluster (Table 2). The cardinality of each cluster indicates the number of similar sequences in that cluster. We report clusters by variant numbering i.e., v1, v2⋯vn; n is the number of clusters or variants for each SARS-CoV-2 protein. The group of similar sequences belonging to a cluster (or variant) for each SARS-CoV-2 protein is reported with accession numbers (Supplementary-1). We draw phylogenetic tree for each SARS-CoV-2 protein taking all distinct variants and report in Supplementary-2. Our analysis shows that distinct variants in ORF1ab, S, N and ORF3a proteins are comparatively higher than other SARS-CoV-2 proteins, signifying that such proteins are highly susceptible.

3.2. Indian vs. world-wide variants

We compare Indian variants with the variants collected from nine (09) major countries such as China (CHN), Bangladesh (BGD), Japan (JPN), Saudi Arabia (SAU), French (FRA), Germany (DEU), Greece (GRC), Italy (ITA), and United States (USA). The protein-specific unique variants observed from all the above countries are reported in Fig. 1(A). We observe a high percentage of unique variants in BGD isolates, followed by Indian isolates. However, the percentage may be an indicator (not conclusive) as the total sample available is non-uniform. We even quantify common variants across nine different countries that are matching with Indian variants is reported in Fig. 1(B). Interestingly, common protein-specific variants are relatively rare while comparing with variants from different countries.

Fig. 1.

Comparison of unique variant among ten different countries. (A) Percentage of unique variant in each SARS-CoV-2 protein. The number at the top of the bar indicates the number of noise-free collected samples; (B) number of common unique variant between India and other country.

3.3. Quantification of observed mutations in SARS-CoV-2 proteins

Among the distinct variants in each SARS-CoV-2 protein, we consider a particular variant as a reference sequence (exactly similar to NC_045512) except ORF1ab protein. We then compare other variants for studying nucleotide level substitutions. The frequency distribution of the number of mutations for each protein is shown in Fig. 2. We observe at least one mutation in case of five proteins (ORF1ab, ORF3a, S, N, M). The average number of mutations per variant for such proteins is relatively higher (Fig. 3). In case of other six proteins (ORF6, ORF7a, ORF7b, ORF8, ORF10, and E), we observe only single mutation in each variant. Upon examining mutations in SARS-CoV-2 proteins, we observe several substitutions, the majority of which are associated with a single variant. The protein wise mutations are highlighted and reported for all the variants associated with more than one samples (Supplementary-3). We list mutations considering only M, N, and S proteins (having mutations in more than one sample). In the case of E protein, only a single mutation is observed in all the variants. In case of accessory proteins, mutations in more than one sample are observed in ORF3a, and ORF8. Most of which are from the non-synonymous category and having more than one sample frequency. We discuss below few top variants and mutations observed in our candidate SARS-CoV-2 proteins.

• •ORF1ab protein: We observe several mutations in ORF1ab because this protein is a polyprotein that consists of sixteen non-structural proteins. We compare all Indian SARS-CoV-2 ORF1ab protein variants with the reference sequence (NC_045512). We observe mutation in 40 variants that are associated with more than one sample. The top non-synonymous mutation ([C14144T, P4715L]) is observed in 233 variants of total 359 samples. Here, the first numbering in bracket refers to the nucleotide mutation position, whereas the second numbering refers to the amino acid substitution position. The majority of the mutations are synonymous. Several non-synonymous mutations are observed associated with five or more samples, which are [A2027C, Q676P], [C18304T, L6102F], [C18890T, T6297I], [G15814A, V5272I], [G10818T, L3606F], [C6047A, T2016K], [C13466T, A4489V], [G4601T, S1534I], [C9173T, T3058I], [C14161A, L4721I]. Further, several synonymous mutations observed in more than one sample are [C2772T, F924F], [C18613T, L6205L], [C2571T, C857C], [G4035T, V1345V], [A16248G, L5416L], [C15060T, N5020N], [C3369T, N1123N], [C3819T, D1273D], [C8517T, S2839S], [C11355T, F3785F].
• •Envelope (E) protein: In the case of E protein, a total 443 (out of 445) variants are observed that are exactly matching with the reference sequence (NC_045512). The only two non-synonymous mutations of single instance are [G184T, V62F] and [G223T, V75F].
• •Membrane (M) protein: A total 224 (out of 457) matching samples (with NC_045512) are found in Indian genome. In M protein, significant synonymous mutations are observed. [C213T, Y71Y] observe in 9 variants of total 223 (≈) 50% samples. In addition, one non-synonymous ([C425T, A142V] in 2 variants of total two samples) and one synonymous ([G429T, V143V] in 2 variants of total four samples) are also observed.
• •Nucleocapsid (N) protein: A total of 204 (out of 455) matching sequences of N protein are observed in Indian samples. We observe mutations in fifty three (53) variants. Each of them is associated with more than one sample. The mutation in the top variant (v2) is [C581T, S194L], which is found in 19 variants of a total of 158 samples. The other important non-synonymous mutations associated in more than one sample are [C581T, S194L], [C38T, P13L], [G605A, S202N], [G608A, R203K], [G609A, R203K], [G610C, G204R], [C614T, T205I], [G578T, S193I], and one synonymous mutations is [G578T, S193I] observed in 2.
• •Spike (S) protein: We observe only 11 samples (out of 436) that are exactly similar to the reference S protein. A total of twenty (20) variants in S protein are found to be associated with mutations in more than one sample. Mutations in each variant show either synonymous or non-synonymous or both the categories. For example, we observe mutations in variant v2 [A1841G, D614G] and [T2367C, Y789Y] that are non-synonymous and synonymous, respectively, and associated with 164 samples. Similarly, the variant (v3) shows two synonymous mutations ([C882T, D294D] and [T2367C, Y789Y]), and two non-synonymous mutations ([G162T, L54F] and [A1841G, D614G]) found in 63 samples. Few findings are consistent with the previously reported results. For example, D614G substitution is observed ≈60% in Indian samples (Saha et al., 2020). In our candidate dataset, we observe D614G substitution in ≈ 93% samples covering 77 variants. The majority of the substitutions are in variant v2, along with a synonymous mutation [T2367C, Y789Y] in the same variant. Few other important mutations are found in five and more samples, which are three non-synonymous mutations ([G162T, L54F], [G1749T, E583D], [G2031T, Q677H]) and three synonymous mutation ([T2367C, Y789Y], [C882T, D294D], [G906T, T302T], [T328C, L110L]).
• •ORF3a protein: In ORF3a protein, we observe only 190 samples (out of 445) in Indian SARS-CoV-2 S proteins, which are exactly similar to the reference sequence. We observe mutations in ORF3a protein of eight (08) variants associated with more than one sample. The top variant is v2 with only non-synonymous mutation [G171T, Q57H] in 17 variants of a total of 234 samples. This non-synonymous mutation (Q57H) is found in Ion channels domain and consistence with previous study (Issa et al., 2020), and shows quite higher percentage ((53%)) in Indian SARS-CoV-2 genome as compared to 17.43% a global study reported in (Issa et al., 2020). Although, another mutation G251V is also found 9.71% of the genomes but, we did not observe this mutation in the Indian SARS-CoV-2 candidate genome. The other observed important mutations associated with more than one sample, where six mutations are non-synonymous ([C121T, L41F], [C277T, H93Y], [G67T, A23S], [C452T, T151I], [G463T, D155Y], [C512T, S171L]), and only one synonymous mutation is [C246T, N82N].
• •ORF6 protein: In the case of ORF6 protein, 457 (out of 459) sequences of SARS-CoV-2 Indian samples are exactly matching with the reference sequence. Similar to E protein, we observe two mutations, each associated with only one sample, one is synonymous ([C12T, L4L]), and the other is non-synonymous ([G39T, E13D]).
• •ORF7a and ORF7b proteins: All the sequences (except two) from the Indian SARS-CoV-2 genome for both the proteins are matched with the reference sequence (NC_045512). Two non-synonymous mutations are observed ORF7a protein associated with two samples each of in a single variant ([C280G, Q94E], [G283A, E95K]). In ORF7b protein, only two non-synonymous are [C92T, S31L] and [G127A, A43T] with an equal number of samples and variants (only 1).
• •ORF8 protein: Majority (423 out of 451) of the sequences are similar with the reference sequence. The top two variants are v2 (non-synonymous mutation: [T251C, L84S]) with sample frequency 19, and v3 (synonymous mutation: [G108T, P36P]) with sample frequency 2.
• •ORF10 protein: In ORF10 protein, 457 out of 460 sequences from the Indian SARS-CoV-2 genome are similar to the reference sequence. The only non-synonymous mutation is [L37F] and synonymous mutation is [C109T] with sample frequency 2 and 1.

Fig. 2.

Distribution of observed number of mutations (x-axis) and relative frequency of number of variants (y-axis) for each SARS-CoV-2 protein. The five proteins ORF1ab, ORF3a, S, N, M are observed multiple mutations in different variants, whereas in six proteins, ORF6, ORF7a, ORF7b, ORF8, ORF10 and E are found exactly a single mutation in each variant.

Fig. 3.

Average number of mutation per variant. Proteins are ranked by avg. mutation, highest (left) to lowest (right).

It can be noted that some of the observed mutations in different variants are common (Supplementary-3). Therefore, with respect to mutation types those variants are highly similar. However, we observe a total of 536 mutated positions located in different SARS-CoV-2 proteins in Indian isolates (Table 3). It is noted that the ORF3a protein shows the highest (≈3.96%) number of mutated locations followed by N protein. We observe a few numbers of mutated locations in E, ORF6, ORF7b and ORF10 proteins.

Table 3.

Number of mutated positions (or locations) in each SARS-CoV-2 protein.

Gene	Gene length	# mutated position	Mutated position (%)
ORF1ab	21,291	328	1.541
E	228	2	0.877
M	669	15	2.242
N	1260	49	3.889
S	3822	83	2.172
ORF3a	828	33	3.986
ORF6	186	2	1.075
ORF7a	336	10	2.976
ORF7b	132	2	1.515
ORF8	366	10	2.732
ORF10	117	2	1.709

3.4. Characterizing the mutations into synonymous and non-synonymous categories

We account for both synonymous and non-synonymous mutations irrespective of any codon positions. Among the observed nucleotide mutations, 541 nucleotide mutations in 536 locations are then characterized in synonymous and non-synonymous categories (see Fig. 4(A)). Overall, percentage of non-synonymous mutation is more (≈62%, count-333) in comparison to synonymous mutations (≈38%, count-208). Observed mutations by the percentage of synonymous and non-synonymous category for all SARS-CoV-2 proteins are shown in Fig. 4(B). Overall, the non-synonymous category percentage is more (except for M protein), where E and ORF7b proteins show 100% non-synonymous mutations.

Fig. 4.

Quantification of synonymous and non-synonymous mutation. (A) Percentage of synonymous vs. non-synonymous mutation type in three codon positions taking all proteins together; (B) percentage of non-synonymous and synonymous mutation type in all SARS-CoV-2 protein.

3.5. Quantifying mutations in three different positions of codon

In case of any coding region mutation may occurs at any three different codon positions. Mutations at the third (3rd) position of the codon are almost synonymous that is the least functionally constrained. In contrast, the majority of the mutations at 1st and 2nd positions of the codon are non-synonymous that alter amino acid. The second codon position is the most functionally constrained as any change to the second codon position causes a non-synonymous change in the coding sequence. We observe mutations in all the three codon positions for ORF1ab, M, N, S, ORF3a, and ORF8 genes (Fig. 5(A)). In, and ORF10 genes, mutations are observed in 1st, 3rd, and 1st codon positions, respectively. We observe mutations in 1st and 2nd positions of ORF7b codons. We do not observe any mutation at the 2nd position of codon. It is worth mentioning that most highly mutated genes (ORF1ab, M, N, S, ORF3a) show a higher percentage of mutations at the third position of the codon, i.e., all these are in the synonymous category.

Fig. 5.

Percentage of mutation in SARS-CoV-2 proteins in each codon position (1st, 2nd and 3rd). (A) Protein-wise in each codon position, and (B) aggregate by all proteins in three different codon positions; (C) overall percentage of synonymous vs. non-synonymous mutation taking all codon positions and proteins.

We account overall mutations that are taking place in all SARS-CoV-2 proteins (Fig. 5(B)). Mutations at 1st and 3nd positions of codons are found almost equal (38%), whereas mutations at 2nd position are comparatively less (24%). More than 90% mutations at 1st codon position are non-synonymous, whereas around 80% mutations at 3rd codon position are synonymous (Fig. 5(C)). Protein-wise the mutations at three different codon positions are reported in Table 4 . In case of non-synonymous mutations, the percentage of mutation at 1st codon position is more (≥50%) for E, ORF10, and ORF7b protein, whereas at 2nd codon position, the mutation percentage is more (≥50%). For ORF8 and ORF7b, and ORF6 highest percentage of mutation occurs at the 3rd codon position.

Table 4.

Percentage of synonymous (syn) and non-synonymous (non-syn) mutation in three different codon positions (CP)-1st/2nd/3rd in each of the SARS-CoV-2 protein.

Protein	CP	Type	Percentage
E	1st	Non-syn	100.00
M	1st	Non-syn	26.67
M	1st	Syn	6.67
N	1st	Non-syn	33.96
N	1st	Syn	1.89
ORF10	1st	Non-syn	50.00
ORF1ab	1st	Non-syn	22.80
ORF1ab	1st	Syn	2.74
ORF3a	1st	Non-syn	33.33
ORF7a	1st	Non-syn	40.00
ORF7b	1st	Non-syn	50.00
ORF8	1st	Non-syn	30.00
S	1st	Non-syn	20.48
S	1st	Syn	1.20
M	2nd	Non-syn	20.00
N	2nd	Non-syn	26.42
ORF1ab	2nd	Non-syn	26.44
ORF3a	2nd	Non-syn	30.30
ORF7a	2nd	Non-syn	20.00
ORF7b	2nd	Non-syn	50.00
ORF8	2nd	Non-syn	60.00
S	2nd	Non-syn	28.92
M	3rd	Syn	46.67
N	3rd	Non-syn	9.43
N	3rd	Syn	28.30
ORF10	3rd	Syn	50.00
ORF1ab	3rd	Non-syn	8.51
ORF1ab	3rd	Syn	39.51
ORF3a	3rd	Non-syn	9.09
ORF3a	3rd	Syn	27.27
ORF6	3rd	Non-syn	50.00
ORF6	3rd	Syn	50.00
ORF7a	3rd	Syn	40.00
ORF8	3rd	Syn	10.00
S	3rd	Non-syn	15.66
S	3rd	Syn	33.73

3.6. Characterizing nucleotide mutation types in non-synonymous category

There are twelve possible nucleotide changes that can occur due to nucleotide mutation (see Methods and Materials). Protein-wise observed mutation counts in three different codon positions are shown in Fig. 6 (A). A majority of the nucleotide mutations are observed in the 1st codon position. Considering three codon positions (Fig. 6(A)), all 12 nucleotide mutation are observed in ORF1ab proteins followed by S proteins. ORF3a and N show comparatively fewer number of mutations (8 and 7 respectively). Similarly, in E, ORF10, and ORF6 proteins only a single mutation is observed.

Fig. 6.

(A) Number of distinct mutation type count shown for non-synonymous category in each protein and each codon position; (B) number of distinct mutation type count shown for non-synonymous category by aggregate all codon positions.

Quantification of nucleotide mutation type shows higher for G > T and C > T, and protein hit count also observer maximum for these two mutation types (Fig. 7(A) and (C)). In terms of percentage (considering all codon positions), the mostly occurring two mutations are C > T (≈32%) and G > T (≈30%). Further, these two mutations (C > T and G > T) are observed in 9 (out of taken 11) SARS-CoV-2 proteins followed by G > A (07) and A > C (05) mutations, respectively. T > A mutation is observed to be rare (only 2). Among the two above mostly occurring mutations, G > T is observed within the top two positions (by percentage) in all three codon positions (Table 5 ), whereas C > T is observed only in 1st and 2nd codon positions. The percentage of abundance of other two important mutations, G > A (in 1st codon position) and A > G (in 2nd codon position), is 17% and 12%, respectively. Further, we observe diversity in different codon positions in individual protein (Table 6 ). For example, G > T is mostly occurring at 2nd codon position (ORFF1ab, S), at the 1st codon position (E, M, N ORF3a), at 3rd codon position (ORF6).

Fig. 7.

The quantification of nucleotide mutation type in non-synonymous category. (A) Percentage of each type of nucleotide mutation; (B) mutation type by associate number of SARS-CoV-2 protein count.

Table 5.

The percentage of nucleotide mutation type for all non-synonymous cases shown for three codon positions independently and arranged by highest to lowest percentage. Mut-type: Mutation type;

Codon position-1st		Codon position-2nd		Codon position-3rd
Mut-type	Percentage	Mut-type	Percentage	Mut-type	Percentage
G>T	33.09	C>T	48.98	G>T	68
C>T	25.00	G>T	16.33	A>C	6
G>A	16.91	A>G	12.24	G>A	6
A>G	6.62	T>C	7.48	G>C	6
A>C	5.88	G>A	6.12	C>A	4
T>C	3.68	C>A	2.72	T>A	4
G>C	2.94	A>C	1.36	T>G	4
C>A	2.21	G>C	1.36	A>T	2
C>G	1.47	T>G	1.36
A>T	0.74	A>T	0.68
T>A	0.74	C>G	0.68
T>G	0.74	T>A	0.68

Table 6.

Percentage of nucleotide mutation type for all non-synonymous cases shown by three codon positions for all proteins. CP: codon position; Mut-type: mutation type.

Protein	CP	Mut-type	Percentage
ORF1ab	2	C>T	23.16
	3	G>T	11.58
	1	C>T	10.53
	1	G>T	10.53
	2	A>G	8.42
	1	G>A	7.37
	2	T>C	4.74
	1	A>G	4.21
	2	G>A	2.63
	2	G>T	2.63
	1	A>C	2.11
	1	T>C	2.11
	2	C>A	1.58
	2	A>C	1.05
	3	C>A	1.05
	3	G>C	1.05
	3	A>C	0.53
	2	A>T	0.53
	1	C>A	0.53
	1	C>G	0.53
	3	G>A	0.53
	1	G>C	0.53
	1	T>A	0.53
	2	T>A	0.53
	1	T>G	0.53
	2	T>G	0.53
E	1	G>T	100.00
M	1	C>T	28.57
	2	C>T	28.57
	1	G>T	28.57
	2	G>T	14.29
N	2	C>T	18.92
	1	G>T	18.92
	1	G>A	10.81
	2	G>T	10.81
	1	C>T	8.11
	2	G>A	5.41
	1	G>C	5.41
	3	G>T	5.41
	3	A>C	2.70
	1	A>G	2.70
	1	C>A	2.70
	3	G>A	2.70
	2	G>C	2.70
	3	G>C	2.70
S	2	C>T	18.52
	2	G>T	18.52
	3	G>T	14.81
	1	G>T	12.96
	1	C>T	5.56
	1	A>C	3.70
	2	A>G	3.70
	3	T>A	3.70
	3	A>C	1.85
	1	A>T	1.85
	1	C>A	1.85
	1	G>A	1.85
	2	G>A	1.85
	3	G>A	1.85
	1	G>C	1.85
	2	G>C	1.85
	1	T>C	1.85
	3	T>G	1.85
ORF3a	1	G>T	20.83
	1	C>T	16.67
	2	C>T	16.67
	2	G>T	12.50
	1	A>C	4.17
	3	A>T	4.17
	2	C>G	4.17
	1	G>A	4.17
	3	G>T	4.17
	2	T>C	4.17
	2	T>G	4.17
	3	T>G	4.17
ORF6	3	G>T	100.00
ORF7a	1	G>A	33.33
	1	C>G	16.67
	1	C>T	16.67
	2	C>T	16.67
	2	G>T	16.67
ORF8	2	C>T	33.33
	1	G>T	22.22
	1	A>C	11.11
	2	C>A	11.11
	2	G>A	11.11
	2	T>C	11.11
ORF7b	2	C>T	50.00
	1	G>A	50.00
ORF10	1	C>T	100.00

3.7. Quantification of non-synonymous amino acid substitutions

As highlighted earlier, there are total 380 amino acid substitutions. Out of 333 non-synonymous substitutions, we observe only 86 distinct substitutions in Indian SARS-CoV-2 genome (Table 7 ).

Table 7.

Amino acid substitution type by associated protein and number of mutated locations in that protein.

Substitution type	Protein (#mutated position)
A>D	ORF1ab-(2)
A>S	M-(1),N-(2),ORF1ab-(4),ORF3a-(3),ORF8-(1),S-(3)
A>T	ORF1ab-(2),ORF7b-(1)
A>V	M-(2),N-(2),ORF1ab-(16),ORF3a-(1),ORF8-(2),S-(4)
C>F	ORF1ab-(1),S-(3)
C>Y	ORF1ab-(1)
D>E	ORF1ab-(1)
D>G	ORF1ab-(4)
D>N	N-(1),ORF1ab-(2)
D>Y	N-(3),ORF1ab-(6),ORF3a-(1),S-(3)
E>D	ORF1ab-(5),ORF6-(1),S-(2)
E>G	ORF1ab-(1)
E>K	ORF1ab-(3),ORF7a-(1)
E>Q	N-(1),ORF1ab-(1),S-(1)
F>L	ORF1ab-(1),S-(1)
G>A	S-(1)
G>C	N-(1),ORF1ab-(3)
G>D	ORF1ab-(3),S-(1)
G>E	ORF8-(1)
G>R	N-(2),ORF1ab-(1)
G>S	N-(1),ORF1ab-(2),S-(1)
G>T	N-(2)
G>V	ORF1ab-(2),ORF3a-(1),ORF7a-(1),S-(1)
G>W	N-(1)
H>Q	ORF3a-(1),S-(1)
H>R	ORF1ab-(2)
H>Y	M-(1),N-(1),ORF1ab-(5),ORF3a-(1),S-(1)
I>K	ORF1ab-(1)
I>L	ORF1ab-(1),ORF8-(1)
I>T	ORF1ab-(4),ORF3a-(1)
K>E	ORF1ab-(1)
K>N	ORF1ab-(7),ORF3a-(1)
K>Q	ORF3a-(1),S-(2)
K>R	ORF1ab-(7),S-(1)
K>T	ORF1ab-(1)
L>F	M-(1),N-(1),ORF10-(1),ORF1ab-(10),ORF3a-(3),ORF7a-(1),S-(2)
L>I	ORF1ab-(1)
L>P	ORF1ab-(2)
L>S	ORF8-(1)
L>V	ORF1ab-(1)
L>W	ORF3a-(1)
M>I	N-(2),ORF1ab-(10),S-(3)
N>D	ORF1ab-(2)
N>H	ORF1ab-(1)
N>K	S-(1)
N>L	ORF1ab-(2)
N>Y	S-(1)
P>A	ORF1ab-(1)
P>L	N-(1),ORF1ab-(7),ORF7a-(1),ORF8-(1)
P>R	ORF3a-(1)
P>S	N-(2),ORF1ab-(5),S-(1)
P>T	N-(1)
Q>E	ORF7a-(1)
Q>H	ORF1ab-(2),S-(4)
Q>K	S-(1)
Q>P	ORF1ab-(1)
Q>R	ORF1ab-(2),S-(1)
R>C	ORF1ab-(2)
R>G	N-(1)
R>I	ORF3a-(1)
R>K	N-(2)
R>L	M-(1),N-(1),ORF1ab-(1),ORF3a-(1)
R>M	S-(1)
R>Q	ORF1ab-(1)
R>S	N-(1)
S>F	ORF1ab-(3),S-(2)
S>G	ORF1ab-(2)
S>I	N-(3),ORF1ab-(1),S-(3)
S>L	N-(1),ORF1ab-(2),ORF3a-(1),ORF7b-(1)
S>N	N-(1)
S>P	ORF1ab-(2)
S>R	ORF1ab-(2)
S>T	ORF1ab-(1)
T>A	ORF1ab-(3)
T>I	N-(3),ORF1ab-(15),ORF3a-(2),S-(4)
T>K	ORF1ab-(1)
T>M	ORF1ab-(1)
T>N	ORF8-(1)
V>A	ORF1ab-(3)
V>F	E-(2),M-(1),ORF1ab-(5),ORF3a-(1)
V>G	ORF1ab-(1)
V>I	ORF1ab-(4),ORF3a-(1),ORF7a-(1)
V>L	ORF1ab-(2),ORF8-(1),S-(1)
W>C	ORF3a-(1)
W>L	S-(2)
Y>H	ORF1ab-(1),S-(1)

We rank amino acid substitutions by the number of substituted positions (Fig. 8(A)). The top substitutions are A > V, which is observed in 27 locations of six (06) different proteins (Fig. 8(B)). The substitution, L > F is observed with maximum hit, occurring in 07 proteins, M (1), N (1), ORF10 (1), ORF1ab (10), ORF3a (3), ORF7a (1), and S (2). Overall it is observed in nineteen (19) different positions of Indian SARS-CoV-2. Similarly, several other important substitutions with regards to number of substituted positions and associated SARS-CoV-2 proteins can be seen from Fig. 8(A) and (B). Further, there are few substitutions, which are observed uniquely in different SARS-CoV-2 proteins. For example, A > D, C > Y, D > E, D > G are observed in ORF1ab protein, G > A, N > K, N > Y, R > M are observed in Spike (S) protein. Several other unique substitutions with their count and type in each SARS-CoV-2 proteins are reported in Fig. 9 and Table 7, respectively. It is to be noted that the highly mutated four proteins are ORF1ab, S, N, and ORF3a. The number of mutations per variant in SARS-CoV-2 proteins of Indian isolates is shown in Fig. 3.

Fig. 8.

(A) The amino acid substitution type observed with more than two mutated positions in SARS-CoV-2 genome. (B) The amino acid substitution type associated with more than two SARS-CoV-2 proteins.

Fig. 9.

The non-synonymous amino acid substitution type count in each of the SARS-CoV-2 protein.

3.8. Functional assessment of non-synonymous amino acid substitutions

Non-synonymous substitutions are vital as they alter the amino acid that impact on the structural and functional imbalance of the target protein. To understanding the functional alteration during non-synonymous substitutions, we use PROVEAN (Choi and Chan, 2015) to predict mutation type whether deleterious or neutral. We calculate ΔΔG values (Capriotti et al., 2005) for predicting the stability variations (increase or decrease or neutral). We report the PROVEAN and ΔΔG scores in Fig. 10 .

Fig. 10.

The non-synonymous amino acid substitution categorization by percentage of deleterious and neutral mutation type predicted by PROVEAN score. (A) Percentage is shown for SARS-CoV-2 proteins taking all codon positions together; (B) percentage is shown for three codon positions in each of SARS-CoV-2 proteins.

It is to be noted that the deleterious percentage is comparatively low for structural proteins (except E) and high for accessory proteins. We predicted a total of 57 (out of 333 non-synonymous substitution) deleterious substitutions as shown in Table 8, Table 9, Table 10 for ORF1ab, structural, and accessory proteins, respectively. All these substitutions are also listed with NCBI protein accession number (Supplementary-4). While considering codon positions of all the deleterious substitutions, we observe that the deleterious substitutions occurs mostly in 2nd codon position (≈51%) followed by 40% and 9% in 1st and 3rd codon positions, respectively (Fig. 10). Moreover, few neutral mutations with a considerable decrease in stability are observed that might impact on protein structural conformation. For example, we observe D614G mutation occurred in the 2nd codon position, which is neutral with a large decrease in stability. This mutation can potentially decrease the structural stability (Maitra et al., 2020). The change in Asp with Gly at this position resulting in the enhancement of local conformational entropy (Ramakrishnan and Ramachandran, 1965). The most frequently observed non-synonymous mutations, Q57H in ORF3a protein and S194L in N protein, occurred in 3rd and 2nd codon positions respectively. In ORF7b, ORF8, and M proteins, deleterious substitutions occur only in 2nd position, whereas in case of ORF6 and ORF10 it is in 3rd and 1st places, respectively. For all other cases, deleterious substitutions are observed either in any two or all three codon positions.

Table 8.

The non-synonymous amino acid substitutions in ORF1ab protein with the predicted PROVEAN score and ΔΔG prediction value.

Substitution	PROVEAN score	Type	ΔΔG prediction	RI	Freq.
G30S	−0.673	Neutral	−1.15	8	1
D33N	−0.733	Neutral	−1.33	4	2
V38F	−0.553	Neutral	−1.48	9	1
G112C	−1.223	Neutral	−1.08	7	1
D147E	−1.123	Neutral	0.01	4	1
V169A	0.027	Neutral	−1.7	8	1
G192D	−1.198	Neutral	−1.23	8	1
L204F	0.327	Neutral	−1.11	8	2
S212L	0.097	Neutral	0.24	1	1
T265I	−0.693	Neutral	−0.67	6	1
T283I	−0.088	Neutral	−0.57	7	2
P309A	−0.135	Neutral	−1.8	8	1
P309L	0.518	Neutral	−0.72	5	1
G327D	−1.072	Neutral	−0.89	5	1
K338R	−0.685	Neutral	0.05	2	1
A339V	−0.465	Neutral	0.07	1	1
E347D	−0.548	Neutral	−0.39	6	1
H417Y	0.379	Neutral	0.26	7	1
S443P	−0.678	Neutral	−0.23	2	1
G519S	−0.633	Neutral	−1.2	8	4
Q575R	−0.331	Neutral	−0.49	6	3
E633D	−0.233	Neutral	−0.37	7	5
E658K	−0.707	Neutral	−0.44	7	1
G662R	−1.425	Neutral	−0.35	7	1
Q676P	−0.531	Neutral	−0.58	7	69
V682L	0.035	Neutral	−1.02	7	1
T882I	−0.691	Neutral	−0.1	2	1
P892S	0.996	Neutral	−1.35	7	1
E940D	0.515	Neutral	−0.41	4	1
G989V	0.35	Neutral	−0.36	5	1
D1036G	−0.887	Neutral	−1.54	8	1
P1054L	−1.268	Neutral	−0.44	0	2
T1055I	−0.496	Neutral	−0.38	2	3
E1126D	−0.535	Neutral	−0.52	6	1
P1158S	−0.909	Neutral	−1.75	9	5
H1160Y	0.734	Neutral	0.11	4	3
V1211F	−0.667	Neutral	−0.7	5	1
E1251K	−0.511	Neutral	−0.7	6	1
A1268T	0.092	Neutral	−0.78	4	2
A1283V	−0.232	Neutral	−0.15	2	2
A1298V	0.22	Neutral	−0.01	1	1
T1429I	0.457	Neutral	−0.5	5	1
A1432V	0.864	Neutral	0.07	2	1
S1534I	0.319	Neutral	0.36	1	7
I1551T	−0.057	Neutral	−1.98	4	1
T1573A	−2.402	Neutral	−1.47	9	1
M1588I	−0.746	Neutral	−0.08	0	2
D1625Y	−2.143	Neutral	0.01	2	2
S1733G	−1.551	Neutral	−1.07	8	2
M1769I	−0.349	Neutral	−0.11	5	3
A1812D	−0.753	Neutral	−0.63	3	6
T1822I	−0.406	Neutral	0.1	0	1
L1853F	−0.808	Neutral	−1.19	6	1
T1854A	−0.326	Neutral	−1.28	8	1
T1854I	−0.193	Neutral	−0.25	3	1
T1874I	−1.364	Neutral	−0.16	3	1
D1939G	−0.936	Neutral	−1.28	6	1
D1940Y	−0.872	Neutral	−0.16	2	1
Q1943H	−0.464	Neutral	−0.78	6	1
K1973R	−0.294	Neutral	−0.37	1	1
S2015R	−0.501	Neutral	−0.17	0	5
T2016K	−0.166	Neutral	−0.86	4	10
K2029E	−0.63	Neutral	−0.5	6	1
K2029N	−0.431	Neutral	−0.64	2	1
P2046L	−1.038	Neutral	−0.65	5	3
T2093I	0.565	Neutral	0.1	4	1
S2103F	−0.372	Neutral	0.24	6	1
L2146P	−1.386	Neutral	−1.61	7	1
S2242P	0.105	Neutral	−0.06	3	1
I2307T	−0.03	Neutral	−2.34	8	1
L2323V	−0.361	Neutral	−1.47	8	1
H2357Y	0.301	Neutral	0.38	7	3
S2488F	2.899	Neutral	−0.05	2	1
K2511N	−0.966	Neutral	−0.38	0	2
H2520R	−0.243	Neutral	−0.29	3	2
A2593V	−1.178	Neutral	−0.24	0	1
A2732D	−3.463	Deleterious	−0.68	6	3
P2739L	−1.595	Neutral	−0.62	4	1
H2831Y	3.17	Neutral	0.27	6	1
A2891V	−0.835	Neutral	0	3	1
D2980G	0.071	Neutral	−1.27	5	1
A2994V	−1.769	Neutral	−0.05	0	3
T3058I	1.463	Neutral	−0.48	4	7
G3072C	−5.058	Deleterious	−1.07	6	2
M3087I	0.614	Neutral	−0.56	5	3
T3150I	0.112	Neutral	−0.43	1	2
S3158G	−0.785	Neutral	−1.35	7	1
L3338F	−3.068	Deleterious	−1.09	6	2
K3353R	−1.343	Neutral	−0.13	1	1
V3377G	−6.124	Deleterious	−2.51	9	1
Q3390R	−0.324	Neutral	−0.33	4	1
N3405L	−4.454	Deleterious	−0.05	0	5
P3447S	−1.913	Neutral	−1.7	9	1
T3453A	−0.882	Neutral	−0.83	7	1
V3475F	−2.291	Neutral	−1.42	9	1
K3499R	−0.421	Neutral	−0.25	2	5
L3606F	−1.432	Neutral	−1	6	14
I3618T	−1.397	Neutral	−1.49	7	1
M3655I	0.174	Neutral	−0.75	7	1
D3681N	−0.466	Neutral	−1.11	7	1
L3711F	−0.348	Neutral	−1.21	7	2
I3731T	−0.744	Neutral	−2.36	9	2
V3759F	−1.765	Neutral	−1.52	9	1
E3909G	−3.759	Deleterious	−1.17	9	1
E3962K	−0.041	Neutral	−0.34	3	1
S3983F	−2.722	Deleterious	−0.34	7	2
R3993C	−6.175	Deleterious	−0.86	5	1
R3993L	−5.422	Deleterious	−0.3	7	1
K4069T	−2.268	Neutral	−0.48	6	1
V4073I	−0.106	Neutral	−0.55	8	1
K4081R	−0.921	Neutral	−0.33	7	1
M4116I	−0.46	Neutral	−0.74	5	1
K4176N	0.651	Neutral	−0.36	3	1
V4181I	−0.046	Neutral	−0.7	7	1
A4271V	−3.278	Deleterious	−0.25	1	1
A4273V	−3.349	Deleterious	−0.23	1	1
K4451N	−0.49	Neutral	−0.65	2	2
K4483N	−1.326	Neutral	−0.42	4	1
A4487V	0.357	Neutral	−0.24	2	1
A4489V	−2.346	Neutral	−0.31	1	10
D4532G	−3.086	Deleterious	−1.08	6	1
A4577V	−1.878	Neutral	−0.17	4	1
M4588I	−1.074	Neutral	−0.81	8	1
I4593L	0.213	Neutral	−0.9	7	1
E4670D	−0.609	Neutral	−0.59	5	2
P4715L	−0.446	Neutral	−0.83	6	359
L4721I	−1.085	Neutral	−1.29	7	7
V4746A	−2.528	Deleterious	−2.01	9	1
M4855I	−1.728	Neutral	−0.81	8	3
C4856F	−0.483	Neutral	−0.21	3	1
L5030F	−2.739	Deleterious	−0.99	7	3
T5035I	−0.622	Neutral	−0.43	2	1
T5036M	−1.529	Neutral	−0.29	2	1
M5060I	−0.117	Neutral	−0.56	7	1
A5091S	−1.821	Neutral	−0.82	9	2
Q5214H	1.016	Neutral	−0.77	6	2
V5272I	−0.551	Neutral	−0.17	2	18
D5285Y	−1.381	Neutral	0.27	3	2
T5300I	−0.542	Neutral	−0.42	4	1
S5305L	−2.332	Neutral	0.22	4	1
P5377S	−0.897	Neutral	−1.73	9	2
H5488Y	0.534	Neutral	0.19	7	1
E5492Q	−2.053	Neutral	−0.7	7	1
G5530C	−2.742	Deleterious	−0.82	4	4
H5569R	1.004	Neutral	−0.1	5	1
V5571F	−1.592	Neutral	−1.62	8	2
Y5577H	−0.845	Neutral	−1.53	7	2
S5583T	−0.858	Neutral	−0.66	4	1
P5624L	−5.36	Deleterious	−0.63	7	2
R5766Q	0.366	Neutral	−0.91	8	1
F5823L	−3.989	Deleterious	−1.17	5	1
A5926S	0.351	Neutral	−1.08	10	1
N5928H	−0.711	Neutral	−0.77	9	1
K5957R	−0.861	Neutral	−0.09	0	1
I5970K	−1.93	Neutral	−2.18	9	1
M5997I	−0.985	Neutral	−0.96	8	1
G6039V	−6.16	Deleterious	−0.14	1	1
A6044V	2.2	Neutral	0.14	2	1
P6065S	0.176	Neutral	−1.67	8	2
L6082F	−0.771	Neutral	−1.04	6	1
R6088C	−5.465	Deleterious	−1.2	7	3
L6102F	−1.397	Neutral	−1.05	4	67
S6180R	−1.897	Neutral	0.16	4	1
A6199S	−2.053	Neutral	−0.62	8	1
D6249Y	0.823	Neutral	−0.16	3	1
K6274N	−0.353	Neutral	−0.18	3	3
T6297I	−0.448	Neutral	−0.8	3	45
N6313D	−3.422	Deleterious	−0.62	7	1
P6368L	−6.762	Deleterious	−0.79	6	1
V6385L	−0.789	Neutral	−1.03	7	3
K6464N	−1.404	Neutral	−0.42	2	1
T6500I	−1.557	Neutral	−0.42	5	1
A6533V	−0.465	Neutral	−0.35	5	3
D6580Y	−0.868	Neutral	−0.59	5	1
G6581D	−2.423	Neutral	−1.12	7	2
A6589V	−0.154	Neutral	−0.17	3	1
V6600A	−2.262	Neutral	−1.84	9	4
L6614F	−1.53	Neutral	−1.28	7	1
A6623T	0.442	Neutral	−0.76	6	1
V6688I	−0.141	Neutral	−0.49	4	1
M6723I	−0.049	Neutral	−0.85	6	1
C6742Y	−0.068	Neutral	−0.18	0	1
D6900Y	−3.735	Deleterious	−0.41	1	1
L6909F	−0.541	Neutral	−1.04	5	1
A6914S	−0.017	Neutral	−1.04	5	2
A6914V	−0.428	Neutral	−0.03	1	1
K6958R	−0.492	Neutral	−0.17	5	2
P7034L	0.713	Neutral	−0.88	5	1
N7083D	−1.153	Neutral	−0.42	2	1

The substitutions with either high PROVEAN score (< − 2.5, type: deleterious) or large increase stability (ΔΔG < − 0.5) or both are shown in bold.

Table 9.

The functional assessment of non-synonymous amino acid substitutions in four structural SARS-CoV-2 proteins (E, M, N, S). The functional assessment of mutation is predicted on utilizing two different measures (PROVEAN score and stability value).

Protein	Substitution	PROVEAN score	Type	ΔΔG prediction	RI	Freq.
E	V62F	0	Neutral	−1.65	8	1
	V75F	−1.414	Neutral	−1.57	9	1
M	A142V	0.18	Neutral	0.25	5	2
	L29F	−1.646	Neutral	−0.91	7	1
	A63V	−1.937	Neutral	0.14	0	1
	A69S	−1.991	Neutral	−0.82	9	1
	V70F	−1.365	Neutral	−1.66	10	1
	R107L	−4.03	Deleterious	−0.33	4	1
	H125Y	0.799	Neutral	0.06	5	1
N	S194L	−4.272	Deleterious	−0.38	2	158
	P13L	−1.23	Neutral	−0.48	3	23
	S202N	−0.404	Neutral	−0.78	0	20
	R203K	−1.604	Neutral	−0.93	6	14
	R203K	−1.604	Neutral	−0.93	6	14
	G204R	−1.656	Neutral	−0.52	5	14
	T205I	−1.562	Neutral	−0.53	3	6
	S193I	−2.755	Deleterious	−0.36	2	4
	G97S	−1.98	Neutral	−1.33	8	3
	A156S	−0.457	Neutral	−0.83	9	3
	P6T	−0.223	Neutral	−1.1	8	2
	S33I	−1.372	Neutral	0.27	6	2
	S180I	−3.465	Deleterious	−0.14	3	2
	M234I	0.044	Neutral	−0.03	1	2
	D22N	−0.541	Neutral	−1.4	6	1
	D22N	−0.541	Neutral	−1.4	6	1
	E31Q	0.054	Neutral	−0.7	5	1
	G34W	−1.609	Neutral	−0.13	2	1
	R92S	−3.718	Deleterious	−1.23	6	1
	G120R	−0.733	Neutral	−0.29	1	1
	A134V	−2.811	Deleterious	−0.12	2	1
	L139F	−0.697	Neutral	−0.85	8	1
	D144Y	−1.764	Neutral	0.2	1	1
	A152S	1.463	Neutral	−0.92	9	1
	R191L	−3.269	Deleterious	−0.58	3	1
	R203G	−3.247	Deleterious	−1.6	7	1
	R203K	−1.604	Neutral	−0.93	6	1
	G204R	−1.656	Neutral	−0.52	5	1
	G204T	−1.76	Neutral	−0.96	7	1
	A218V	0.171	Neutral	0.21	1	1
	M234I	0.044	Neutral	−0.03	1	1
	G236C	−2.269	Neutral	−0.27	5	1
	H300Y	−1.577	Neutral	0.46	5	1
	P302S	−4.043	Deleterious	−1.3	7	1
	P344S	−4.031	Deleterious	−1.46	8	1
	D348Y	−0.588	Neutral	−0.41	2	1
	T362I	−1.722	Neutral	−0.35	3	1
	T393I	−0.613	Neutral	0.1	2	1
S	D614G	0.598	Neutral	−0.93	3	405
	L54F	−0.435	Neutral	−1.14	4	80
	E583D	−0.819	Neutral	−0.44	3	14
	R78M	0.986	Neutral	−0.84	7	12
	T572I	−0.649	Neutral	0	3	10
	Q677H	0.002	Neutral	−0.67	5	5
	L5F	−1.126	Neutral	−0.98	3	3
	Q690H	−0.796	Neutral	−0.86	6	3
	S12F	−0.65	Neutral	0.14	2	2
	W152L	−0.159	Neutral	−0.89	7	2
	S155I	−0.503	Neutral	0	4	2
	M177I	0.579	Neutral	−0.61	5	2
	G181A	0.396	Neutral	−0.58	1	2
	W258L	−1.084	Neutral	−0.65	7	2
	A706S	0.183	Neutral	−0.85	9	2
	A879S	−0.361	Neutral	0.54	7	2
	H1083Q	−1.006	Neutral	−0.34	5	2
	C1243F	−4.53	Deleterious	−0.09	2	2
	F2L	−0.902	Neutral	−1.22	9	1
	S13I	−1.187	Neutral	0.27	0	1
	Y28H	−0.443	Neutral	−1.38	4	1
	G35V	−2.112	Neutral	−0.68	7	1
	T76I	−0.115	Neutral	−0.72	6	1
	K97Q	−0.113	Neutral	−0.92	7	1
	N148Y	−0.177	Neutral	0.1	4	1
	M153I	0.244	Neutral	−0.98	6	1
	E156D	0.958	Neutral	−0.52	4	1
	S162I	0.231	Neutral	0.02	1	1
	Q173H	−0.299	Neutral	−1.02	7	1
	S255F	−0.423	Neutral	−0.03	3	1
	G261S	0.485	Neutral	−1.13	8	1
	A262S	0.154	Neutral	−0.64	9	1
	Q271R	−0.48	Neutral	−0.27	5	1
	C301F	−8.689	Deleterious	0.2	4	1
	E471Q	0.445	Neutral	−0.59	7	1
	D574Y	0.858	Neutral	0.36	2	1
	Q613H	−0.917	Neutral	−0.86	6	1
	H655Y	−0.814	Neutral	0.08	4	1
	A688V	0.498	Neutral	−0.37	5	1
	A701V	0.597	Neutral	−0.25	4	1
	M731I	−0.598	Neutral	−0.25	3	1
	K795Q	0.072	Neutral	−0.61	3	1
	P809S	1.024	Neutral	−1.55	8	1
	T827I	−0.378	Neutral	−0.45	6	1
	A892V	−1.901	Neutral	0.2	1	1
	A930V	−3.727	Deleterious	−0.2	3	1
	T1077I	−1.511	Neutral	−0.13	1	1
	V1104L	−0.604	Neutral	−0.7	1	1
	D1153Y	−3.275	Deleterious	−1.52	2	1
	K1181R	−0.522	Neutral	−0.48	7	1
	N1187K	−0.467	Neutral	−0.29	4	1
	Q1201K	1.409	Neutral	−0.29	3	1
	C1250F	−5.057	Deleterious	−0.09	2	1
	D1259Y	3.924	Neutral	−0.21	3	1

The substitutions with either high PROVEAN score (< − 2.5, type: deleterious) or large increase stability (ΔΔG < − 0.5) or both are shown in bold.

Table 10.

The functional assessment of non-synonymous amino acid substitutions in six SARS-CoV-2 accessories proteins (ORF3a, ORF6, ORF7a, ORF7b, ORF8, ORF10). The functional assessment of mutation is predicted on utilizing two different measures (PROVEAN score and stability value).

Protein	Substitution	PROVEAN score	Type	ΔΔG prediction	RI	Freq.
ORF3a	G18V	−1.571	Neutral	−0.28	6	1
	K21Q	0.657	Neutral	−0.47	1	1
	A23S	−1.638	Neutral	−0.86	9	2
	I35T	−2.619	Deleterious	−2.39	9	1
	L41F	−2.724	Deleterious	−1.08	7	4
	P42R	−5.495	Deleterious	−0.96	7	1
	V50I	−0.657	Neutral	−0.84	8	1
	L53F	−3.962	Deleterious	−1.09	7	1
	A54S	−1.638	Neutral	−0.6	8	1
	Q57H	−3.286	Deleterious	−0.9	7	234
	K66N	3.486	Neutral	−0.16	1	1
	R68I	−1.562	Neutral	0.17	3	1
	V77F	2.638	Neutral	−1.37	8	1
	L86W	−3.943	Deleterious	−1.13	1	1
	H93Y	−3.943	Deleterious	0.3	6	3
	A103V	−2.876	Deleterious	0.2	5	1
	L108F	−3.4	Deleterious	−1.24	6	1
	W131C	−7.752	Deleterious	−1.29	8	1
	R134L	−1.543	Neutral	−0.47	9	1
	A143S	0.724	Neutral	−0.95	9	1
	T151I	−4.886	Deleterious	−0.29	0	2
	D155Y	−6.829	Deleterious	0.21	0	2
	S171L	−2.238	Neutral	−0.22	0	2
	T175I	2.562	Neutral	−0.04	4	1
ORF6	E13D	−2.786	Deleterious	−0.24	4	1
ORF7a	Q94E	−1	Neutral	−0.24	2	2
	E95K	−2.614	Deleterious	−0.6	8	2
	G38V	−6.526	Deleterious	−0.4	4	1
	P45L	−10	Deleterious	−0.7	4	1
	V71I	−0.667	Neutral	−0.24	5	1
	L116F	−1.263	Neutral	−0.85	7	1
ORF7b	S31L	−6	Deleterious	0.23	1	1
	A43T	0	Neutral	−0.44	5	1
ORF8	L84S	2.333	Neutral	−2.29	8	19
	G8E	−3.056	Deleterious	−0.6	1	1
	T12N	−1.056	Neutral	−0.71	1	1
	A14S	0.833	Neutral	−0.47	6	1
	A51V	−1.222	Neutral	−0.06	2	1
	V62L	−0.722	Neutral	−0.8	5	1
	A65V	1.222	Neutral	0.02	1	1
	P85L	−8.778	Deleterious	−0.73	7	1
	I121L	−0.278	Neutral	−0.79	5	1
ORF10	L37F	NA	Deleterious	−0.99	6	1

The substitutions with either high PROVEAN score (< − 2.5, type: deleterious) or large increase stability (ΔΔG < − 0.5) or both are shown in bold.

We also predict the stability impact of single point mutations, where most of the substitutions show a large decrease in stability. We find a total of 32 single point deleterious substitutions (7 from structural proteins) out of total 57 with large decrease in stability (ΔΔG < − 0.5). We highlight all the functional domains of all such non-synonymous deleterious substitutions in Table 11 . Additionally, we study the changes in physicochemical properties during such substitutions. A few numbers of substitution leads to change physicochemical property both in hydropathy class and side-chain structural classes (Table 11).

Table 11.

The 57 deleterious amino acid substitutions in different SARS-CoV-2 proteins highlighted with the putative functional domain and physicochemical property changes. The mutations with large decrease stability (ΔΔG < − 0.5) are shown in bold.

Protein	Substitution	Putative functional domain	Hydropathy change	Chemical property change
ORF1ab	A2732D	NSP3	Hydrophobic to charge	Aliphatic to acidic
	G3072C	NSP4	Hydrophilic (unchanged)	Aliphatic to sulfur containing
	L3338F	NSP5 (3CLpro)	Hydrophobic (unchanged)	Aliphatic to aromatic
	V3377G	NSP5 (3CLpro)	Hydrophobic to hydrophilic	Aliphatic to aliphatic
	N3405L	NSP5 (3CLpro)	Hydrophilic to hydrophobic	Acidic amide to aliphatic
	E3909G	NSP7	Charge to hydrophilic	Acidic to aliphatic
	S3983F	NSP8	Hydrophilic to hydrophobic	Hydroxyl containing to aromatic
	R3993C	NSP8	Charge to hydrophilic	Basic to sulfur containing
	R3993L	NSP8	Charge to hydrophilic	Basic to aliphatic
	A4271V	NSP10	Hydrophobic (unchanged)	Aliphatic (unchanged)
	A4273V	NSP10	Hydrophobic (unchanged)	Aliphatic (unchanged)
	D4532G	NSP12 (RdRp)	Charge to hydrophilic	Acidic to aliphatic
	V4746A	NSP12 (RdRp)	Hydrophobic (unchanged)	Aliphatic (unchanged)
	L5030F	NSP12 (RdRp)	Hydrophobic (unchanged)	Aliphatic to aromatic
	G5530C	NSP13 (helicase)	Hydrophilic (unchanged)	Aliphatic to sulfur containing
	P5624L	NSP13 (helicase)	Hydrophobic (unchanged)	Cyclic to aliphatic
	F5823L	NSP13 (helicase)	Hydrophobic (unchanged)	Aromatic to aliphatic
	G6039V	NSP14 (exonuclease)	Hydrophilic to hydrophobic	Aliphatic to aliphatic
	R6088C	NSP14 (exonuclease)	Charge to hydrophilic	Basic to sulfur containing
	N6313D	NSP14 (exonuclease)	Hydrophilic to charge	Acidic amide to acidic
	P6368L	NSP14 (exonuclease)	Hydrophobic (unchanged)	Cyclic to aliphatic
	D6900Y	NSP16	Charge to hydrophobic	Acidic to aromatic
M	R107L	Topological domain	Charge to hydrophobic	Basic to aliphatic
N	R92S	NTD	Charge to hydrophilic	Basic to hydroxyl containing
	A134V	NTD	Hydrophobic (unchanged)	Aliphatic (unchanged)
	S180I	SR-rich linker	Hydrophilic (unchanged)	Hydroxyl containing to aliphatic
	R191L	SR-rich linker	Charge to hydrophobic	Basic to aliphatic
	S193I	SR-rich linker	Hydrophilic to hydrophobic	Hydroxyl containing to aliphatic
	S194L	SR-rich linker	Hydrophilic to hydrophobic	Hydroxyl containing to aliphatic
	R203G	SR-rich linker	Charge to hydrophilic	Basic to aliphatic
	P302S	CTD	Hydrophobic to hydrophilic	Cyclic to hydroxyl containing
	P344S	CTD	Hydrophobic to hydrophilic	Cyclic to hydroxyl containing
S	C301F	S1 (N-terminal)	Hydrophilic (unchanged)	Sulfur containing to aromatic
	A930V	S2 (HR-1)	Hydrophobic (unchanged)	Aliphatic (unchanged)
	D1153Y	S2 (between HR1 and HR2)	Charge to hydrophilic	Acidic to aromatic
	C1243F	S2 (cytoplasm domain)	Hydrophilic to hydrophobic	Sulfur containing to aromatic
	C1250F	S2 (cytoplasm domain)	Hydrophilic to hydrophobic	Sulfur containing to aromatic
ORF3a	I35T		Hydrophobic to hydrophilic	Aliphatic to hydroxyl containing
	L41F	TM-1	Hydrophobic (unchanged)	Aliphatic to aromatic
	P42R	TM-1	Hydrophobic to charge	Cyclic to basic
	L53F	TM-1	Hydrophobic (unchanged)	Aliphatic to aromatic
	Q57H	TM-1	Hydrophilic to charge	Acidic amide to basic
	L86W	TM-2	Hydrophobic (unchanged)	Aliphatic to aromatic
	H93Y	Ion channels	Charge to hydrophilic	Basic to aromatic
	A103V	Ion channels	Hydrophobic (unchanged)	Aliphatic (unchanged)
	L108F	Ion channels	Hydrophobic (unchanged)	Aliphatic to aromatic
	W131C	Ion channels	Hydrophobic to hydrophilic	Aromatic to sulfur containing
	T151I	C-terminal	Hydrophilic to hydrophobic	Hydroxyl containing to aliphatic
	D155Y	C-terminal	Charge to hydrophilic	Acidic to aromatic
ORF6	E13D		Charge (unchanged)	Acidic (unchanged)
ORF7a	G38V	Luminal domain	Hydrophilic (unchanged)	Aliphatic to aliphatic
	P45L	Luminal domain	Hydrophobic (unchanged)	Cyclic to aliphatic
	E95K	Luminal domain	Charge (unchanged)	Acidic to basic
ORF7b	S31L		Hydrophilic to hydrophobic	Hydroxyl containing to aliphatic
ORF8	G8E	N-terminal (hydrophobic region)	Hydrophilic to charge	Aliphatic to acidic
	P85L		Hydrophobic (unchanged)	Cyclic to aliphatic
ORF10	L37F		Hydrophobic (unchanged)	Aliphatic to aromatic

The ORF1ab protein consists of several non-structural polyproteins (NSP1-NSP16). A few deleterious substitutions are detected in putative functional domains of those polyproteins. The 3-chymotrypsin-like cysteine protease (3CLpro) and RNA-dependent RNA polymerase (RdRp) regions located in NSP3 and NSP12 polyproteins, respectively. It’s playing a major role in anti viral drug discovery for SARS-CoV-2 and other coronavirus diseases (ul Qamar et al., 2020; Anand et al., 2003; Calligari et al., 2020). So, the mutations detected in these functional domains might impact protein functions and stability. Three deleterious substitutions are also detected both in 3CLpro and RdRp region of ORF1ab polyprotein. The few deleterious substitutions with large decrease in stability changes are detected in other two important functional domains, namely helicase in NSP13 (Chen et al., 2020; Yu et al., 2012), and exonuclease in NSP14 (Yuen et al., 2020). These are also investigated to inhibit coronavirus (Chen et al., 2020; Shannon et al., 2020; Yu et al., 2012).

The Membrane (M) protein is one of the most abundant structural proteins among coronaviruses protein and has an interaction role with other structural proteins (He et al., 2004; Naskalska et al., 2019). A single deleterious substitution is observed in Topological domain (Bianchi et al., 2020).

N protein contains two distinct RNA-binding domains- the N-terminal domain (NTD, 44–179 residues) and the C-terminal domain (CTD, 247–363 residues) (Zeng et al., 2020), responsible for RNA binding and oligomerization, respectively. These two regions are connected by an intrinsically disordered central Ser/Arg (SR)-rich linker (Kang et al., 2020), which is responsible for primary phosphorylation. The study on the Nucleocapsid protein of other coronaviruses, several residues of N-terminal domain, is associated with RNA binding and virus infectivity (Saikatendu et al., 2007; Tan et al., 2006; Grossoehme et al., 2009). Among the observed seven deleterious substitutions in N protein, we observe 2 in NTD, 5 in SR-rich linker, and 2 in CTD functional domain (Kang et al., 2020).

The S1 subunit (residues: 14–685) and the S2 subunit (residues: 686–1273) in Spike protein regions are responsible for receptor binding and membrane fusion, respectively (Huang et al., 2020). The N-terminal domain (residues: 14–305) belongs to the S1 subunit. The S2 subunit consists of several sub-domains, including heptapeptide repeat sequence 1 (HR1) (residues: 912–984), HR2 (residues: 1163–1213), cytoplasm domain (residues: 1237–1273) (Xia et al., 2020). We observe five deleterious substitution in Spike protein, one each in S1 (N-terminal), S2 (HR-1), and S2 subunit in between HR1 and HR2. Two deleterious substitution occur in S2 subunit (Cytoplasm domain) (Huang et al., 2020; Walls et al., 2020).

SARS-CoV has three major transmembrane domains: (i) Transmembrane domain 1 (TM-1) (approx. residues: 34–56), (ii) transmembrane domain 2 (TM-2) (approx. residues: 77–99), and (iii) transmembrane domain 3 (TM-3) (approx. residues: 103–125) available mostly in ORF3a and C-terminal domain with about 160 amino acid residues (Hofman, 1993; Zeng et al., 2004). In connection with approximate residues of SARS-CoV, We observe four mutations in TM-1 and one in TM-2 domains of SARS-CoV-2 ORF3a protein. Four important deleterious substitutions are observed in the Ion channels domain (Domain II, residues: 91–133) (Issa et al., 2020), which is linked to its pro-apoptotic function as observed for other SARS-coronavirus (Chan et al., 2009; Lu et al., 2010). One of the observed mutations is W131C, located in Cysteine rich region (cysteine-rich region overlaps the third membrane-spanning domain) of ORF3a protein (Zeng et al., 2004). This mutation further increases the Cysteine residue in that region that may alter interchain disulfide linkages with the Spike protein of other viral structural proteins. Additionally, two mutations in the C-terminal domain are observed between the last two Cysteine residues (Zeng et al., 2004).

Both the proteins ORF6 and ORF8 do not have any trans-membrane regions, but ORF8 has an hydrophobic signal peptide (residues: 1–15) and chain (residues: 61–121) (Alam et al., 2020). However, they play significant roles in innate immune suppression during viral infection, regulation of molecular functions, virus growth, replication, and host interactions (Alam et al., 2020; Li et al., 2020; Mohammad et al., 2020). A single deleterious substitution (E13D) is found in ORF6 protein and two mutations in ORF8 protein, where one in the Hydrophobic region (G8E) (Alam et al., 2020; Mohammad et al., 2020).

The domain of ORF7a protein of Indian SARS-CoV-2 consists of seven (07) β strands (Alam et al., 2020). A similar result is reported for ORF7a protein of SARS-CoV in (Nelson et al., 2005; Bartlam et al., 2007). It consists of N-terminal signal peptide (residues: 1–15), luminal domain (16–96), transmembrane segment (residues: 97–117), and a 5 residue cytoplasmic tail. Considering the similar organizational domains of SARS-CoV with SARS-CoV-2, three deleterious substitutions are identified in the luminal domain, two of them (G38V and P45L) are located before and after the 3rd β strand.

4. Conclusion

In this study, we thoroughly investigated and characterized mutations observed in Indian SARS-CoV-2 genome. We reported variants and mutations observed in all the SARS-CoV-2 proteins belong to both synonymous and non-synonymous categories. We highlighted position-specific mutations in the codons. Non-synonymous amino acid substitutions are analyzed further to predict the functional stability of the proteins.

Our study reported a total of 536 mutated positions in the coding region of SARS-CoV-2 proteins. The ORF3a happens to be the mostly mutated protein (≈4% of total length), followed by three structural proteins (N, M, S). However, both in ORF3a and N proteins, we observed fewer mutation types compared to ORF1ab and S. The number of variants and mutations per variant observed to be maximum for ORF1ab followed by Spike protein. Interestingly, counts for non-synonymous mutations are higher compared to synonymous mutations (except for M protein). Mutations in E and ORF7b proteins are all non-synonymous.

Our analysis further reveals that most of the deleterious substitutions with decrease in stability occur in the 2nd position (codon) and putative functional domains. Higher quantity of single point mutation, G > T, is observed both in 1st and 3rd positions in the codon, whereas mutation C > T, shows maximum occurrence in 2nd codon position. The conclusion drawn purely based on computational analysis, needs experimental confirmation. Though we restricted our current study on Indian isolates, it may easily be extended to any other strains. Overall analysis might help in better understanding of the possible role in virulence, infectivity, and virus release in SARS-CoV-2. A further comparative study on the significant mutations observed in Indian isolates may be performed with the strains collected from the rest of the world.

The following are the supplementary data related to this article.

Supplementary-1

Accession numbers of SARS-CoV-2 collected sequences from Indian isolates.

Supplementary-2

The polygenetic tree for each of the SARS-CoV-2 proteins.

Supplementary-3

The variant with frequency and respective nucleotide mutation and type, amino acid change, and number of mutations for each of the SARS-CoV-2 proteins.

Supplementary-4

The observed non-synonymous substitution with the protein accession number in each of the SARS-CoV-2 proteins.

CRediT authorship contribution statement

Jayanta Kumar Das: Conceptualization, Data curation, Methodology, Software, Formal analysis, Visualization, Writing - original draft, review & editing. Antara Sengupta: Methodology, Software, Validation, Visualization, Review & editing. Pabitra Pal Choudhury: Review & editing. Swarup Roy: Conceptualization, Supervision, Visualization, Review & editing.

Declaration of competing interest

The authors declare that they have no competing interests.