Abstract

   Lung cancer is the leading cause of mortality from cancer worldwide.
   Lung adenocarcinoma (LUAD) is a type of non-small cell lung cancer
   (NSCLC) with highest prevalence. Kinesins a class of motor proteins are
   shown to be involved in carcinogenesis. We conducted expression, stage
   plot and survival analyses on kinesin superfamily (KIF) and scrutinized
   the key prognostic kinesins. Genomic alterations of these kinesins were
   studied thereafter via cBioPortal. A protein–protein interaction
   network (PPIN) of selected kinesins and 50 closest altering genes was
   constructed followed by gene ontology (GO) term and pathway enrichment
   analyses. Multivariate survival analysis based on CpG methylation of
   selected kinesins was performed. Lastly, we conducted tumor immune
   infiltration analysis. Our results found KIF11/15/18B/20A/2C/4A/C1 to
   be significantly upregulated and correlated with poor survival in LUAD
   patients. These genes also showed to be highly associated with cell
   cycle. Out of our seven selected kinesins, KIFC1 showed the highest
   genomic alteration with highest number of CpG methylation. Also, CpG
   island (CGI) cg24827036 was discovered to be linked to LUAD prognosis.
   Therefore, we deduced that reducing the expression of KIFC1 could be a
   feasible treatment strategy and that it can be a wonderful individual
   prognostic biomarker. CGI cg24827036 can also be used as a therapy site
   in addition to being a great prognostic biomarker.

   Subject terms: Computational biology and bioinformatics, Genetics,
   Immunology, Structural biology, Systems biology

Introduction

   Lung cancer (LC) is a prevalent and deadly disease that ranks first
   among cancers in terms of death and 2nd most diagnosed cancer in both
   genders globally^[46]1. The etiological and molecular heterogeneity of
   LC contributes greatly to treatment failure and adverse survival
   outcomes^[47]2,[48]3. Most LCs diagnosed are malignant epithelial
   tumours, which can be further classified as small-cell lung carcinoma
   (SCLC) or non-small cell lung carcinoma (NSCLC). NSCLC accounts for
   85–90% of lung malignancies, with lung adenocarcinoma (LUAD) and lung
   squamous cell carcinoma (LUSC) being the most frequent
   subtypes^[49]4,[50]5. LUAD and LUSC can be classified into four stages,
   referred to as
   [MATH: <mi mathvariant="normal">I</mi> :MATH]
   ,
   [MATH: <mi mathvariant="normal">II</mi> :MATH]
   ,
   [MATH: <mi mathvariant="normal">III</mi> :MATH]
   , and
   [MATH: <mi mathvariant="normal">IV</mi> :MATH]
   , as per the tumor node metastasis (TNM) taxonomy^[51]6. The early,
   non-metastatic stage is referred to as stage
   [MATH: <mi mathvariant="normal">I</mi> :MATH]
   . Stages
   [MATH: <mi mathvariant="normal">II</mi> :MATH]
   and
   [MATH: <mi mathvariant="normal">III</mi> :MATH]
   typically represent the intermediate, regional lymphatic metastatic
   phases, with stage
   [MATH: <mi mathvariant="normal">III</mi> :MATH]
   exhibiting more significant metastasis in the lymphatic region than
   stage
   [MATH: <mi mathvariant="normal">II</mi> :MATH]
   . Meanwhile, stage
   [MATH: <mi mathvariant="normal">IV</mi> :MATH]
   often denotes a late stage with distant metastases^[52]6.

   Despite evidence that smoking increases the risk of LUAD, it is
   currently the most common subgroup of LC among non-smokers and
   women^[53]7,[54]8. Patients with LUAD typically have a poor prognosis
   and frequently show local progression or metastasis when
   diagnosed^[55]9. However, LUSC is more prevalent in men than in women
   and has been strongly linked to smoking^[56]10. Although chemotherapy,
   radiation, and targeted medicines are widely employed, therapeutic
   resistance to these treatments is a primary cause of treatment failure.
   Understanding the underlying molecular pathways of carcinogenesis is
   thus critical for developing effective LC therapies.

   Human kinesin superfamily members (KIFs) consist of
   [MATH: <mrow><mn>14</mn></mrow> :MATH]
   kinesin family members, kinesin-1 to kinesin-14, according to the
   standardized nomenclature adopted by the kinesin research group^[57]11.
   There are
   [MATH: <mrow><mn>45</mn></mrow> :MATH]
   members in the KIFs superfamily, including
   [MATH: <mrow><mn>39</mn></mrow> :MATH]
   N-kinesins, three M-kinesins, and three C-kinesins^[58]12. KIF proteins
   are a family of motor proteins that move molecules and depend on
   microtubules. They have ATPase activity as well as motion
   characteristics. They bind to microtubules and then move along the
   microtubules, carrying protein complexes, organelles, and messenger
   RNAs (mRNAs)^[59]12–[60]14. In recent years, it has come to light that
   several KIFs contribute uniquely to the process of mitosis, also known
   as cell division, by taking part in the motion of chromosomes and
   spindles^[61]15,[62]16. Additionally, individual kinesins are also
   essential for a number of other cellular processes, such as endocytosis
   and transcytosis, intracellular transport^[63]14.

   Mitosis, the process by which eukaryotic cells divide, creates two
   daughter cells with approximately equal amounts of the cell's nucleus,
   cytoplasm, organelles, and membrane. It is possible that mistakes in
   this process could lead to the death of cell, abnormalities (including
   deletion of gene, translocation of chromosome, or the duplication of
   chromosomes), and even cancer. Since mitosis is so intricately
   controlled therefore, any alteration or changes in KIF expression or
   function could potentially cause cancer. Kinesins and motor proteins
   with abnormal expression are crucial mitotic process regulators
   and potential targets in human malignancies^[64]17–[65]19. Human cancer
   is a genetic disorder characterized by uncontrolled cell development,
   hence inhibiting kinesins may provide a unique approach to managing
   this disease.

   Thus, identifying anomalous kinesin gene expression could be utilized
   as a biomarker for early tumor diagnosis and targeting kinesins could
   also be a novel approach for cancer therapy. Therefore, in the current
   study, we conducted a comprehensive bioinformatics analysis to identify
   the key kinesins influencing the prognosis of LUAD cancer patients. We
   performed expression and stage plot analyses of KIFs across the cancer
   genome atlas (TCGA)-LUAD patient samples and reported only significant
   ones. Next, we proceeded with overall survival (OS) analysis followed
   by mutational, enrichment, and protein–protein interaction network
   (PPIN) analyses. At last, we obtained KIFC1 as final prognostic
   biomarker responsible for LUAD pathogenesis. KIFC1 can be further used
   for early detection of LUAD patients and targeted therapy or
   personalized medicine.

Materials and methods

Kinesins expression and stage analysis across LUAD cohort

   Gene expression profiling interactive analysis v2 (GEPIA 2) web-based
   tool^[66]20 ([67]http://gepia2.cancer-pku.cn/) was accessed for
   comparing the relative mRNA expression level of all kinesin family
   members across TCGA-LUAD cohort and matched TCGA normal and GTEx data.
   The expression values from GEPIA were already transformed into
   [MATH: <mrow><msub><mi
   mathvariant="normal">log</mi><mn>2</mn></msub><mrow><mo
   stretchy="false">(</mo><mi
   mathvariant="normal">TPM</mi><mo>+</mo><mn>1</mn><mo
   stretchy="false">)</mo></mrow></mrow> :MATH]
   values followed by differential analysis. Pathological stage plot
   analysis was also done with GEPIA 2 to investigate the kinesin family
   members' expression with respect to different pathological stages in
   LUAD. The threshold used in GEPIA for mRNA expression level comparison
   across LUAD and normal samples were as follows:
   [MATH: <mrow><mi>p</mi><mi
   mathvariant="normal">value</mi><mo><</mo><mn>0.05</mn></mrow> :MATH]
   and
   [MATH: <mrow><mfenced close="|" open="|"><msub><mi
   mathvariant="normal">log</mi><mn>2</mn></msub><mrow><mo
   stretchy="false">(</mo><mrow><mi mathvariant="normal">fold</mi><mi
   mathvariant="normal">change</mi></mrow><mo
   stretchy="false">)</mo></mrow></mfenced><mo>></mo><mn>1</mn></mrow>
   :MATH]
   . Kinesins statistically significant in both expression and stage plot
   analyses were selected for further analyses.

Prognostic analysis of kinesins across LUAD cohort

   Kaplan–Meier (KM) plotter^[68]21,[69]22
   ([70]https://kmplot.com/analysis/) was queried for prognostic analysis
   of kinesins having significance in expression and stage plot analyses.
   We generated KM plots of only those kinesins which showed significant
   OS across LUAD patient samples. The microarray LUAD patients were
   bifurcated into higher and lower expression groups based on their
   median values. The redundant samples were removed in the quality
   control section, and biased arrays were excluded. Hazard ratio (HR)
   with the corresponding 95% confidence interval (CI),
   [MATH: <mrow><mtext>logrank</mtext><mspace
   width="0.277778em"></mspace><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext></mrow> :MATH]
   and median survival were calculated.
   [MATH: <mrow><mtext>logrank</mtext><mspace
   width="0.277778em"></mspace><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo><</mo><mn>0.05</mn>
   </mrow> :MATH]
   was considered as a statistically significant threshold for assessing
   the prognosis of kinesins between two expression groups.

Validation of prognostic kinesins using cBioPortal

   We queried the cBioPortal for Cancer Genomics^[71]23
   ([72]https://www.cbioportal.org/) for investigating the mutations and
   putative copy number alterations (CNAs) of prognostically significant
   kinesins. The LUAD dataset (TCGA, Firehose Legacy) was chosen to
   perform our analysis.

Validation of prognostic kinesins using GEO and correlation analysis

   We queried the NCBI- GEO^[73]24 ([74]https://www.ncbi.nlm.nih.gov/geo/)
   using “LUAD” and “Lung Adenocarcinoma” as suitable keywords for
   extracting LUAD-associated mRNA expression profile. All the search
   results were further trimmed down in accordance with the following
   inclusion criteria: (1) the samples present in dataset(s) must belong
   to ‘Homo Sapiens’; (2) dataset(s) type must be ‘expression profiling by
   array’; (3) both preprocessed and raw files of the dataset(s) must be
   available; (4) the dataset(s) submission date to GEO must be within
   last
   [MATH: <mrow><mn>10</mn></mrow> :MATH]
   years (i.e. 2012–2022); (5) the dataset(s) must be comprising both
   tumor and healthy control tissue samples; (6) the dataset(s) must
   comprise at least
   [MATH: <mrow><mn>25</mn></mrow> :MATH]
   samples. Any abstracts, case reports, review-based articles,
   cell-line-based experimental study designs, and studies devoid of
   healthy controls or non-human samples were excluded. Sequential steps
   of batch correction, probe ID to gene mapping, and duplicacy removal
   were performed as discussed previously^[75]25. The DEGs were screened
   corresponding to a Benjamini-Hochberg (BH)—p value < 0.0 and
   [MATH: <mrow><mfenced close="|"
   open="|"><mrow><msub><mtext>log</mtext><mn>2</mn></msub><mfenced
   close=")" open="("><mrow><mtext>fold
   change</mtext></mrow></mfenced></mrow></mfenced><mo>></mo><mn>0.5</mn><
   /mrow> :MATH]
   utilizing limma^[76]26. The presence of key prognostic kinesins was
   checked in the DEGs list. Next, we accessed GEPIA 2 to perform pairwise
   correlation analysis of key prognostic kinesins across TCGA-LUAD and
   normal patients. p value < 0.05 was considered as the cutoff for
   statistical significance.

PPIN construction and enrichment analysis

   A PPIN was constructed between the prognostically significant kinesins
   and top
   [MATH: <mrow><mn>50</mn></mrow> :MATH]
   frequently altered genes corresponding to a default confidence (i.e.,
   interaction score
   [MATH: <mrow><mo>></mo><mn>0.4</mn></mrow> :MATH]
   ) using Search Tool for the Retrieval of Interacting Genes (STRING)
   v11.5 web-based tool^[77]27 ([78]https://string-db.org/) and visualized
   via Cytoscape v3.9.1^[79]28. Top
   [MATH: <mrow><mn>10</mn></mrow> :MATH]
   significant (i.e.,
   [MATH: <mrow><mi mathvariant="normal">p</mi><mo>-</mo><mi
   mathvariant="normal">value</mi><mo><</mo><mn>0.05</mn></mrow> :MATH]
   ) pathway and gene ontology (GO) terms for the constructed PPIN items
   were compiled using Enrichr web server^[80]29
   ([81]https://maayanlab.cloud/Enrichr). Kyoto Encyclopedia of Genes and
   Genomes (KEGG)^[82]30–[83]32, GO-Biological Process (BP), GO-Molecular
   Function (MF), and GO-Cellular Compartment (CC) libraries were used for
   pathway and GO terms.

Tumor infiltration analysis

   We looked into the relationship between mRNA expression levels of
   prognostically significant kinesins with tumor-infiltrating immune
   cells such as B cells,
   [MATH: <msup><mrow><mi
   mathvariant="normal">CD</mi><mn>8</mn></mrow><mo>+</mo></msup> :MATH]
   T cell, macrophage, and neutrophils across TCGA-LUAD patients using
   TIMER 2.0^[84]33 ([85]http://timer.cistrome.org/). To assess the
   statistical significance, Spearman correlation was used.

Methylation analysis

   Prognostic analysis of single CpG methylation of selected genes of
   kinesin family in LUAD patients was conducted using MethSurv^[86]34
   ([87]https://biit.cs.ut.ee/methsurv), a web tool for multivariate
   survival analysis based on CpG methylation data.

Results

Kinesins expression and stage plot analysis across LUAD cohort

   All kinesins' relative mRNA expression distribution across TCGA-LUAD
   cohort (
   [MATH: <mrow><mn>483</mn></mrow> :MATH]
   tumor and
   [MATH: <mrow><mn>347</mn></mrow> :MATH]
   normal) was compiled utilizing GEPIA. KIF11, KIF12, KIF15, KIF23,
   KIF18B, KIF20A, KIF2C, KIF4A, KIFC1 expression levels were
   significantly upregulated while KIF17, KIF26A, KIF1C expressions were
   significantly downregulated in tumor samples as shown by the
   box-and-whisker plots in Fig. [88]1A–L. All these significantly
   expressed kinesins were carried further to stage plot analysis. The
   pathological sub-stage analysis as shown by violin plots in
   Fig. [89]2A–H revealed that overexpressed levels of KIF11, KIF15,
   KIF23, KIF18B, KIF20A, KIF2C, KIF4A, KIFC1 significantly correlated
   with advanced TNM stages across TCGA-LUAD cohort.

Figure 1.

   [90]Figure 1
   [91]Open in a new tab

   Box-and-whisker plots displaying the relative mRNA expression levels of
   (A) KIF11, (B) KIF12, (C) KIF15, (D) KIF17, (E) KIF18B, (F) KIF20A, (G)
   KIF23, (H) KIF26A, (I) KIF1C, (J) KIF2C, (K) KIF4A, (L) KIFC1 across
   TCGA-LUAD and normal samples. Grey-and red-colored box areas signify
   normal and tumor patient samples. The top and bottom of the boxes
   signify 75th and 25th percentile of distribution. Horizontal lines
   within the boxes represent the median values while minimum and maximum
   values label the axes endpoints. *
   [MATH: <mrow><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo><</mo><mn>0.05</mn>
   </mrow> :MATH]
   .

Figure 2.

   [92]Figure 2
   [93]Open in a new tab

   Violin plots displaying association between significant TNM sub-stages
   and mRNA expression levels of (A) KIF11 (B) KIF15, (C) KIF18B, (D)
   KIF20A, (E) KIF23, (F) KIF2C, (G) KIF4A, (H) KIFC1 across TCGA-LUAD
   cohort. The black-colored vertical bars and white-colored dots signify
   interquartile ranges and median, respectively. The ordinate and
   abscissa depict expression levels of these genes and various stages.
   Distribution density is represented by the width of turquoise-colored
   shapes, respectively.

Prognostic analysis of kinesins across LUAD cohort

   Using KM plotter, prognostic analysis was performed on KIF11, KIF15,
   KIF23, KIF18B, KIF20A, KIF2C, KIF4A, KIFC1 to determine the correlation
   between their mRNA expression levels and risk of
   [MATH: <mrow><mn>513</mn></mrow> :MATH]
   LUAD patient samples. The KM plots as shown in Fig. [94]3A–G revealed
   significantly poor OS of LUAD patients when mRNA expression levels of
   KIF11, KIF15, KIF18B, KIF20A, KIF2C, KIF4A, and KIFC1 were high. The
   low and high expression cohort median survival time, HR,
   [MATH: <mrow><mn>95</mn><mrow><mo>%</mo><mi
   mathvariant="normal">CI</mi></mrow></mrow> :MATH]
   , and
   [MATH: <mrow><mrow><mi mathvariant="normal">logrank</mi><mi
   mathvariant="normal">p</mi></mrow><mo>-</mo><mi
   mathvariant="normal">value</mi></mrow> :MATH]
   of each kinesin is detailed in Supplementary Table [95]S1,
   respectively.

Figure 3.

   [96]Figure 3
   [97]Open in a new tab

   KM plots showing the OS of (A) KIF11 (B) KIF15, (C) KIF18B, (D) KIF20A,
   (E) KIF2C, (F) KIF4A, (G) KIFC1 across LUAD microarray cohort. Red and
   black colors signify higher and lower expression groups.

Validation of key prognostic kinesins using cBioPortal

   We used cBioPortal to validate the specific genetic modifications
   associated with key prognostic kinesins (i.e., KIF11, KIF15, KIF18B,
   KIF20A, KIF2C, KIF4A, KIFC1) across LUAD dataset (TCGA, Firehose
   legacy) comprising
   [MATH: <mrow><mn>584</mn></mrow> :MATH]
   tumor patient samples. OncoPrint results for these queried genes as
   represented in Fig. [98]4 revealed genetic alterations in
   [MATH: <mrow><mn>8</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>49</mn><mo stretchy="false">/</mo><mn>584</mn></mrow>
   :MATH]
   ) patient samples. As observed, KIFC1 showed maximum mutation frequency
   (
   [MATH: <mrow><mn>2.3</mn><mo>%</mo></mrow> :MATH]
   ) as compared to others. The cancer type summary analysis revealed the
   overall alteration frequency of these genes as shown in Supplementary
   Figure [99]S1. We observed
   [MATH: <mrow><mn>0.78</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>4</mn><mo stretchy="false">/</mo><mn>516</mn></mrow>
   :MATH]
   cases) missense mutation and
   [MATH: <mrow><mn>0.39</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>2</mn><mo stretchy="false">/</mo><mn>516</mn></mrow>
   :MATH]
   cases) deep deletion in case of KIF11. In case of KIF15, we observed
   [MATH: <mrow><mn>0.58</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>3</mn><mo stretchy="false">/</mo><mn>516</mn></mrow>
   :MATH]
   cases) missense mutation and
   [MATH: <mrow><mn>0.19</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>1</mn><mo stretchy="false">/</mo><mn>516</mn></mrow>
   :MATH]
   case) deep deletion. In case of KIF18B, we observed
   [MATH: <mrow><mn>0.78</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>4</mn><mo stretchy="false">/</mo><mn>516</mn></mrow>
   :MATH]
   cases) amplification and
   [MATH: <mrow><mn>0.58</mn><mo>%</mo></mrow> :MATH]
   (
   [MATH: <mrow><mn>3</mn><mo stretchy="false">/</mo><mn>516</mn></mrow>
   :MATH]
   cases) missense mutation. In case of KIF20A, we observed
   [MATH: <mrow><mn>0.58</mn><mo>%</mo></mrow> :MATH]
   (3/516 cases) missense mutation, 0.39% (2/516 cases) deep deletion, and
   0.19% (1/516 case) amplification. In case of KIF2C, we observed 0.19%
   (1/516 case) truncating mutation and 1.55% (8/516 cases) amplification.
   In case of KIF4A, we observed 0.19% (1/516 case) deep deletion, 0.39%
   (2/516 cases) amplification, and 1.36% (7/516 cases) missense mutation.
   In case of KIFC1, we observed 1.55% (8/516 cases) amplification and
   0.78% (4/516 cases) missense mutation.

Figure 4.

   [100]Figure 4
   [101]Open in a new tab

   OncoPrint summarizing genomic alterations of key prognostic kinesins
   across TCGA-LUAD cohort comprising 584 patient samples. The bottom row
   represents frequency of genomic alterations in KIF11, KIF15, KIF18B,
   KIF20A, KIF2C, KIF4A, KIFC1 with red, blue, green, orange, and grey
   bars signifying amplifications, deep deletions, missense, splice, and
   truncating mutations, respectively. First, second, third, fourth, and
   fifth rows depicts the clinical annotation bars such as profiled in
   putative copy-number alterations from GISTIC, mutation spectrum, sex,
   tissue source site, and mutation count, respectively.

Validation using GEO and correlation analysis

   As per the specified inclusion and exclusion criteria we chose
   [102]GSE43458 (30 healthy control + 80 tumor tissues) and
   [103]GSE116959 (11 healthy control + 57 tumor tissues) LUAD-associated
   mRNA expression profiles. A total of 2861 and 5128 DEGs were screened
   corresponding to [104]GSE43458 and [105]GSE116959 as per the specified
   threshold. The lists of DEGs are shown in Supplementary Tables [106]S2
   and [107]S3. All the key prognostic kinesins (i.e., KIF11, KIF15,
   KIF18B, KIF20A, KIF2C, KIF4A, and KIFC1) were present in the DEGs lists
   of both datasets, thus confirming their validation in external GEO
   datasets. Strikingly, all the prognostic kinesins were upregulated
   among DEGs list and matched with the primary results obtained form
   GEPIA 2. Scatterplots showing pairwise correlations among these key
   prognostic kinesins are demonstrated in Supplementary Figures
   [108]S2–[109]S5. Significantly highest correlation between KIF4A and
   KIF2C (
   [MATH: <mrow><mtext>R</mtext><mo>=</mo><mn>0.95</mn></mrow> :MATH]
   ,
   [MATH: <mrow><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo>=</mo><mn>6.9</mn><
   mo>×</mo><msup><mn>10</mn><mrow><mo>-</mo><mn>269</mn></mrow></msup></m
   row> :MATH]
   ) was observed.

PPIN construction and enrichment analysis

   Our PPIN comprised a total of 57 nodes and 1455 edges as shown in
   Fig. [110]5. Within PPIN, degree, betweenness, and closeness values
   ranged from 4 to 56, 0.07 to 40.44, and 0.51 to 1. The average degree,
   betweenness, and closeness of PPIN were 51.05, 4.94, and 0.931.
   Topological/centrality measures like node degree, betweenness,
   closeness, clustering coefficient, neighborhood connectivity, and
   average shortest path length of PPIN are demonstrated in Supplementary
   Figure [111]S6. Subsequently, we performed pathway and GO term
   enrichment analysis on key prognostic kinesins and associated top 50
   frequently altered genes. Barplots showing top 10 significantly
   enriched pathway and GO terms is shown in Fig. [112]6. The most
   significant pathway, GO-BP, GO-MF, GO-CC terms were cell cycle (
   [MATH: <mrow><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo>=</mo><mn>4.8</mn><
   mo>×</mo><msup><mn>10</mn><mrow><mo>-</mo><mn>14</mn></mrow></msup></mr
   ow> :MATH]
   ), microtubule cytoskeleton organization involved in mitosis (
   [MATH: <mrow><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo>=</mo><mn>1.44</mn>
   <mo>×</mo><msup><mn>10</mn><mrow><mo>-</mo><mn>38</mn></mrow></msup></m
   row> :MATH]
   ), microtubule binding (
   [MATH: <mrow><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo>=</mo><mn>1.54</mn>
   <mo>×</mo><msup><mn>10</mn><mrow><mo>-</mo><mn>21</mn></mrow></msup></m
   row> :MATH]
   ), spindle (
   [MATH: <mrow><mi>p</mi><mspace
   width="0.277778em"></mspace><mtext>value</mtext><mo>=</mo><mn>5.46</mn>
   <mo>×</mo><msup><mn>10</mn><mrow><mo>-</mo><mn>36</mn></mrow></msup></m
   row> :MATH]
   ). Most number of genes corresponding to pathway, GO-BP, GO-MF, GO-CC
   terms were 11, 25, 18, 38 for cell cycle, mitotic spindle organization,
   microtubule binding, intracellular membrane-bounded organelle.

Figure 5.

   [113]Figure 5
   [114]Open in a new tab

   PPIN comprising 57 nodes and 1455 edges. Magenta-colored nodes
   represent prognostic kinesins and green-colored nodes represent top 50
   frequently altered genes.

Figure 6.

   [115]Figure 6
   [116]Open in a new tab

   Barplots showing top 10 significantly enriched (A) pathways, (B) GO-BP,
   (C) GO-MF, (D) GO-CC terms with respect to p values. The color of bars
   varies in accordance with p values with red signifying lowest p values
   and green signifying highest p values. Asterisk signs represent the
   terms are also significant according to FDR.

Tumor infiltration analysis

   Correlation of KIF11, KIF15, KIF18B, KIF20A, KIF2C, KIF4A, KIFC1 mRNA
   expression levels with tumor purity and infiltrating levels of
   neutrophils, macrophages, B cells, and CD8^+T cell across TCGA-LUAD
   cohort are shown by scatterplots in Fig. [117]7. KIF11 displayed
   significant positive correlations with infiltrating levels of CD8^+T
   cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.18</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>5.85</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>5</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.231</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.03</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>7</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.154</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>6.09</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>4</mn></mrow></msup></mrow> :MATH]
   ). KIF15 displayed significant positive correlations with infiltrating
   levels of
   [MATH: <mrow><mtext>CD</mtext><msup><mn>8</mn><mo>+</mo></msup></mrow>
   :MATH]
   T cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.21</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.69</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>6</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.277</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>3.88</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>10</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.155</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>5.39</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>4</mn></mrow></msup></mrow> :MATH]
   ). KIF18B displayed significant positive correlations with infiltrating
   levels of CD8^+T cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.135</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.61</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>3</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.214</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.65</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>6</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.106</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.91</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>2</mn></mrow></msup></mrow> :MATH]
   ). KIF20A displayed significant positive correlations with infiltrating
   levels of CD8^+T cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.115</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.06</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>2</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.229</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.61</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>7</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.108</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.69</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>2</mn></mrow></msup></mrow> :MATH]
   ). KIF2C displayed significant positive correlations with infiltrating
   levels of CD8^+T cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.161</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>3.19</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>4</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.207</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>3.64</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>6</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.146</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.15</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>3</mn></mrow></msup></mrow> :MATH]
   ). KIF4A displayed significant positive correlations with infiltrating
   levels of
   [MATH: <mrow><mtext>CD</mtext><msup><mn>8</mn><mo>+</mo></msup></mrow>
   :MATH]
   T cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.192</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.71</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>5</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.262</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>3.38</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>9</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.195</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.31</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>5</mn></mrow></msup></mrow> :MATH]
   ). KIFC1 displayed significant positive correlations with infiltrating
   levels of CD8^+T cell (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.137</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.28</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>3</mn></mrow></msup></mrow> :MATH]
   ), neutrophils (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.193</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.60</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>5</mn></mrow></msup></mrow> :MATH]
   ), and macrophages (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.128</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>4.27</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>3</mn></mrow></msup></mrow> :MATH]
   ). KIF11 (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.24</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>6.58</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>8</mn></mrow></msup></mrow> :MATH]
   ), KIF15 (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.188</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.65</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>5</mn></mrow></msup></mrow> :MATH]
   ), KIF18B (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.17</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.55</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>4</mn></mrow></msup></mrow> :MATH]
   ), KIF20A (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.218</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>1.03</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>6</mn></mrow></msup></mrow> :MATH]
   ), KIF2C (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.221</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>6.92</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>7</mn></mrow></msup></mrow> :MATH]
   ), KIF4A (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.22</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>7.76</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>7</mn></mrow></msup></mrow> :MATH]
   ), KIFC1 (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mo>-</mo><mn>0.164</mn></mrow>
   :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>2.58</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>4</mn></mrow></msup></mrow> :MATH]
   ) showed significant negative correlations with infiltrating levels of
   B cells. In addition, KIF11 (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.028</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>5.36</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ), KIF15 (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.016</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>7.21</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ), KIF18B (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.002</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>9.58</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ), KIF20A (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.019</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>6.80</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ), KIF2C (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.007</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>8.77</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ), KIF4A (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.01</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>8.24</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ), KIFC1 (
   [MATH: <mrow><mtext>r</mtext><mo>=</mo><mn>0.031</mn></mrow> :MATH]
   ,
   [MATH:
   <mrow><mi>p</mi><mo>=</mo><mn>4.97</mn><mo>×</mo><msup><mn>10</mn><mrow
   ><mo>-</mo><mn>1</mn></mrow></msup></mrow> :MATH]
   ) showed nonsignificant positive correlations with tumor purity across
   TCGA-LUAD cohort.

Figure 7.

   [118]Figure 7

   [119]Figure 7
   [120]Open in a new tab

   Scatterplots showing significant correlations of (A) KIF11, (B) KIF15,
   (C) KIF18B, (D) KIF20A, (E) KIF2C, (F) KIF4A, (G) KIFC1 with
   infiltrating levels of CD8^+T cell, B cells, neutrophils, and
   macrophages across TCGA-LUAD cohort. Spearman’s correlation value and
   estimated statistical significance were shown as the legends for each
   scatter plot.

Prognostic analysis based on single CpG methylation of selected kinesins in
LUAD patients

   We obtained the heatmaps of DNA methylation of selected kinesins using
   MethSurv. Among which cg04344917 CpG island (CGI) of KIF11, cg09053247
   CGI of KIF15, cg01838385 CGI of KIF18B, cg07632946 CGI of KIF20A,
   cg20487572 CGI of KIF2C, cg27286863 CGI of KIF4A, cg2390442 CGI of
   KIFC1 showed the highest methylation levels (Fig. [121]8). Furthermore,
   we studied KM plots which revealed that cg24827036 CGI of KIFC1 were
   significantly associated with survival of LUAD patients (Fig. [122]9).
   A total of 461 patients were split into higher and lower expression
   groups. Higher methylated expression of KIFC1 worsened the OS of LUAD
   patients.

Figure 8.

   [123]Figure 8

   [124]Figure 8
   [125]Open in a new tab

   Heatmaps of CpG methylation levels of (A) KIF11, (B) KIF15, (C) KIF18B,
   (D) KIF20A, (E) KIF4A, (F) KIF2C, (G) KIFC1 across LUAD patients. Rows
   indicates the CpGs and columns indicates the patients. Methylation
   levels (1 = fully methylated; 0 = fully unmethylated) are shown as a
   continuous variable from red to blue color, high expression to low
   expression. Various colorful side boxes were used to represent the
   event, relation to UCSC_CpG_island and UCSC_refGene_Group.

Figure 9.

   Figure 9
   [126]Open in a new tab

   KM plot showing single CpG methylation of KIFC1 across LUAD patients.
   It’s location relative to CpG island, gene sub-region, CpG ID, and gene
   ID are also shown. Red and blue colors signify higher and lower
   expression groups.

Discussion

   LUAD’s malignancy results in high morbidity and fatality
   rate^[127]35,[128]36. Despite advances in surgery, radiation, and
   chemotherapy, which have improved tumor patients' clinical prognosis
   and survival^[129]37, LUAD is still hard to treat because scientists
   don't fully understand the molecular mechanisms and basic signaling
   pathways in how LC works. It is expected that molecule-targeted therapy
   will be a revolutionary treatment technique for solid tumors, however,
   its efficacy and advantages remain restricted^[130]38,[131]39. Because
   of chemoresistance and recurrence, the currently available therapeutic
   choices are limited. Therefore, a new and effective molecular target
   must be identified to cure LUAD.

   Members of the KIF gene family are mostly found in eukaryotic cells,
   namely microtubules. Experiments conducted in vitro have shown that the
   transport of proteins occurs in only one direction, along the
   microtubule's negative pole and in the direction of the positive pole.
   Therefore, the genes that make up the KIF family are responsible for
   controlling the movement of mass proteins both inside of cells and
   outside of cells. This control encompasses a variety of functions, such
   as moving organelles and vesicles that contain material and taking part
   in the process of cell mitosis^[132]15,[133]40,[134]41. There have been
   reports that several genes in the kinesin family are linked to
   different kinds of cancer^[135]42–[136]44. KIF family member genes have
   been demonstrated in various cancer types to establish their prognostic
   and diagnostic capacities. In our current study, we performed
   expression analysis of kinesin family in LUAD which revealed
   overexpression of KIF11/12/15/23/18B/20A/2C/4A/C1 in tumor samples
   whereas KIF17/26A/1C were underexpressed in LUAD. Furthermore, we also
   studied mRNA expression based on cancer stage which showed
   overexpression of KIF11/15/23/18B/20A/2C/4A/C1 in tumor tissues.
   Furthermore, we evaluated the prognostic value of selected kinesins in
   LUAD patients. Our results showed that an increased
   KIF11/15/18B/20A/2C/4A/C1 expression is associated with poor OS in LUAD
   patients. So, by targeting these kinesins and decreasing their effect
   can be of therapeutic importance and patients’ survival can be
   increased.

   Our findings corroborate with multiple previous findings that showed
   overexpression of KIF11, KIF15, KIF18B, KIF20A, KIF2C, and KIF4A in
   LUAD tissues and when LUAD patients have high expression of these KIFs,
   their chances of survival are lower^[137]38,[138]45–[139]49. Next, we
   studied the genomic alterations of key kinesins which showed the
   highest alteration in KIFC1 (2.3%) as amplification being the most
   prominent type of alteration. Following that we constructed a PPIN of
   key kinesins and top 50 frequently altered genes and performed
   enrichment analysis. Our results showed high enrichment of kinesins in
   cell cycle and oocyte meiosis pathway, in biological processes named
   microtubule cytoskeleton organization involved in mitosis and mitotic
   spindle organization, microtubule binding and microtubule motor
   activity molecular functions and spindle and microtubule cytoskeleton
   cellular components.

   For cells to divide and multiply, they go through a series of events
   known as the cell cycle, and abnormalities in the control of the genes
   involved in the cell cycle have been linked to the development of
   tumours. Mutations in upstream signal transduction pathways or genetic
   abnormalities within genes that encode cell cycle proteins cause
   cancer^[140]50. Our result showed high enrichment of kinesins in cell
   cycle processes hence kinesins are involved in controlling these
   processes somewhat and regulating LUAD.

   The other two typically active mechanisms in LC were oocyte meiosis and
   progesterone-mediated oocyte maturation. One cycle of DNA replication
   in meiosis is followed by two cycles of chromosomal segregation
   (Meiosis I and Meiosis II). Normally, oocytes are stopped during the G2
   stage of meiosis I. Progesterone exposure releases them from this
   natural lock, allowing the two meiotic division cycles to resume and
   the oocyte to mature^[141]51,[142]52. So, it makes sense that
   dysregulations in oocyte maturation and meiosis would impact the cell
   cycle process, and further cell cycle changes would impact normal
   bodily functions, increasing the likelihood that one would develop
   cancer. So, cell cycle, progesterone-mediated oocyte maturation, and
   oocyte meiosis play prominent roles in the progression of LUAD^[143]53
   and these two processes come under top 10 in the pathway enrichment
   analysis we did in our study showing the importance of kinesins in
   controlling these pathways in LUAD Microtubules are
   [MATH: <mi mathvariant="normal">α</mi> :MATH]
   - and
   [MATH: <mi mathvariant="normal">β</mi> :MATH]
   -tubulin heterodimers polymers. They exhibit highly dynamic behaviour,
   continuously engaging through processes of polymerization and
   de-polymerization, as well as lengthening and shortening. The
   fundamental components of the cytoskeleton are actin, intermediate
   filaments, and microtubules. They are crucial for various cell
   processes, including mitosis, the movement of vesicles and organelles
   inside cells, cell signaling, migration through cilia and flagella,
   cell shape and morphology^[144]54. So it can be said that any
   alteration in the function of kinesin from normal can impact these
   important pathways involved in LUAD progression.

   We also performed tumor immune infiltration analysis on our key
   kinesins as tumor-infiltrating immune cells are critical components of
   the tumour microenvironment (TME), influencing tumor growth and
   survival depending upon their type and interaction LC clearly displays
   an invasion of a wide variety of immune cell types comprising
   neutrophils, natural killer (NK) cells, macrophages, dendritic cells, T
   cells & B cells^[145]55. These cells perform multiple purposes and
   combine or oppose one another, producing the LC TME. Neutrophils make
   up 50–70% of all white blood cells in the bloodstream and serve as the
   body's initial defence against infections. Neutrophils have been found
   to enhance tumor growth through various clinically relevant mechanisms.
   Tumor growth, angiogenesis, tumor cell migration, and metastasis are
   all facilitated by neutrophils. Still, a subclass of TANs known as N1
   can have anticancer properties^[146]56. Our data revealed the highest
   correlation of our key kinesins with infiltration abundances of
   neutrophils in LUAD patients. CD8^+T cell, also known as cytotoxic T
   lymphocytes, play a crucial role in mounting an efficient antitumor
   response. These cells can identify specific tumor-associated antigens
   (TAA) that are presented on major histocompatibility complex (MHC)
   class I molecules on the surface of cancer cells. Furthermore, they
   possess the ability to eliminate cancer cells directly^[147]57. Our
   study found that infiltrations of CD8^+T cell and neutrophil correlated
   most with KIF15 expression levels. Also, B cell and macrophage
   infiltrations correlated most with KIF11 and KIF4A expression levels.

   Cancer is often caused by the inactivation of several tumor-suppressor
   genes through point mutations and deletion of chromosomes^[148]58.
   Recent research has shown that epigenetic changes are key to cancer
   development. Many genes have CGIs in their promoter regions, and
   abnormal methylation of these sites in cancer leads to transcriptional
   suppression. Epigenetic alterations are passed down through cell
   division, resulting in gene activity change but no changes in the
   sequence of DNA^[149]59,[150]60. Changes in DNA methylation patterns
   are a key feature of many types of cancer, including LC.

   So further we conducted a single CpG methylation-based prognostic
   analysis on key kinesins which showed that CpG methylation in KIFC1 was
   associated with poor prognosis in LUAD patients. KIFC1 is believed to
   be an oncogene in various types of cancers as it plays a crucial role
   in clustering multiple centrosomes to sustain tumor
   survival^[151]61,[152]62.

Conclusions

   Our research revealed a significant function for the kinesin family in
   initiating and progressing LUAD. KIF11/15/18B/20A/2C/4A/C1 mRNA
   expression levels were significantly upregulated and correlated with
   poor OS across LUAD patients. They were highly associated with cell
   cycle. Our results revealed the highest genomic alteration in KIFC1
   with highest number of CpG methylation. cg24827036 CGI of KIFC1 was
   associated with poor prognosis across LUAD. We concluded that KIFC1 can
   be a great individual prognostic biomarker, and that inhibiting its
   expression could be a potential therapeutic approach. Additionally, CpG
   island cg24827036 can serve as a great prognostic biomarker and
   treatment site.

Supplementary Information

   [153]Supplementary Information 1.^ (1.8MB, docx)
   [154]Supplementary Table S2.^ (184.6KB, xlsx)
   [155]Supplementary Table S3.^ (319.3KB, xlsx)

Acknowledgements