Abstract The research aims to understand Alzheimer’s genetic and immune landscapes using the amalgamation of three technologies: artificial intelligence (GenAI), integrative bioinformatics, and single-cell analysis. First, the study aims to identify and characterize the significant genes associated with Alzheimer’s disease (AD) using three GenAI models (GPT‑4o, Gemini model, and DeepSeek). After the genes were accumulated from GenAI models, 27 genes associated with AD were recoded. Furthermore, they were analyzed using integrative bioinformatics methods. Similarly, the immune landscape of AD using single-cell analysis was also explored, which reveals a high percentage of effector CD8^+ T cells (33.42%) and naive T cells (45.95%). The single-cell study found that effector memory T cells have two subsets. It also found that the macrophage population has started to spread and dendritic cells have decreased in Alzheimer’s patients. The single-cell gene expression study reveals the top ten highly expressed genes (NDUFV2, CAT, MRPS34, PBX3, THOC2, CCDC57, PBXIP1, SDHAF3, PPP4C, and MAP3K8). The clonal frequency indicates that CD8^+ T and naive T cell populations show the highest clonal frequency in healthy and AD individuals and are further noted them in the clonotype cell proportion study. Following our GenAI and single-cell profiling strategy, future studies will help in quickly understanding the genetic and immune basis of many diseases. Keywords: Bioinformatics, GenAI, single-cell analysis, Alzheimer’s disease, genetic and immune landscape Graphical abstract graphic file with name fx1.jpg [31]Open in a new tab __________________________________________________________________ Das and colleagues demonstrated Alzheimer’s genetic and immune landscapes using GenAI, instigative bioinformatics, and single-cell analysis. This study informed 27 significant genes associated with Alzheimer’s disease using three GenAI models, which were further analyzed using instigative bioinformatics. Conversely, the single-cell study illustrated the immune landscape of Alzheimer’s disease. Introduction Over the past few years, there has been an artificial intelligence (AI) boom worldwide. Researchers are using machine learning (ML) to deep learning (DL) models to process extensive amounts of data. Recently, Generative AI (GenAI) has been a type of AI that can assemble new content like text, videos, music, and images.[32]^1^,[33]^2^,[34]^3 These tools, also known as large language models (LLMs), show new options in several fields of biology, including medical science.[35]^4^,[36]^5^,[37]^6 With the help of NLP (natural language processing), these tools can design contextually and coherently relevant textual content.[38]^4^,[39]^7^,[40]^8 It attempts to explore different areas of medical fields, including drug discovery, nucleic acid research, clinical medicine and their applications.[41]^4^,[42]^9^,[43]^10^,[44]^11^,[45]^12^,[46]^13 LLMs are one of the remarkable technologies that have developed recently due to the faster progression of AI, and they are a noteworthy achievement. In functioning, the tokenization concept of LLM experiences training on considerable volumes of textual data. LLMs can perform significantly, and there is now extreme interest among researchers. They are using the model in research very significantly.[47]^5 LLM-enabled chatbots can identify questions and provide answers automatically. OpenAI’s LLM product ChatGPT uses generative pre-trained transformer (GPT) architecture and it was released on November 30, 2022.[48]^14 After it was released, it became famous, and the GPT model users were recorded to be about 3.5 million within a few months.[49]^15 OpenAI released the GPT-4 model, an MLLM (multimodal LLM) that can be trained with image, audio, or video.[50]^16 Similarly, Google DeepMind developed the Gemini MLLM model that can use 1 million tokens.[51]^17 Recently, China developed another LLM model, DeepSeek, which delights scientists who use it in scientific research. It uses different parameters and 1.8T tokens.[52]^18^,[53]^19^,[54]^20^,[55]^21 The GenAI or LLMs can retrieve and sort the data. It is used in biological research.[56]^22 Previously, while using LLM of LMM, we used prompt engineering-related GenAI models (LLM or MLLM) to identify the antibody escape mutations in NTD (N-terminal domain) and RBD (receptor binding domain) regions of S-protein (Spike-protein) in SARS-CoV-2.[57]^5 Therefore, GenAI models are essential for molecular biology research for mutation identification and other feature identification. Understanding the genetic landscape helps us comprehend the disease’s regulation, which is crucial in developing and progressing many diseases. Understanding the cellular factors that influence processes that potentially lead to genetic to epigenetic changes will help us understand the total landscape of the disease.[58]^23^,[59]^24 Therefore, it is essential to understand a disease’s genetic landscape. At the same time, it is necessary to identify genes and genetic variations that help in disease development and progression.[60]^25^,[61]^26^,[62]^27^,[63]^28^,[64]^29 Understanding the druggable genes will assist us in drug development.[65]^30 It will also help us understand the therapeutic target and aid in therapeutic development. Therefore, it is imperative to understand the genetic landscape of a disease. However, people are trying to understand the genetic landscape of different diseases, primarily neurological diseases.[66]^25^,[67]^31 Van Cauwenberghe tried to comprehend Alzheimer’s genetic landscape, and it has massive clinical implications, which help in the clinical diagnosis of Alzheimer’s disease (AD) and its genetic risk.[68]^32 However, one critical problem is that the genetic landscape of disease is associated with many genes.[69]^33^,[70]^34 Therefore, it is a significant challenge to identify crucial genes, and GenAI might be an essential tool for identifying causative genes or significant genes in disease. AD is a slow and progressive a neurodegenerative disease that has affected millions of people worldwide. The early stages of the disease involve minor loss of memory and subsequent progression into severe disfigurement of broad cognitive functions. It is also the most common form of dementia.[71]^35^,[72]^36^,[73]^37 The pathobiology of this disease shows a wide range of clinical manifestations, which include neuronal and synaptic loss, glial cell activation, paired helical filament, β-amyloid accumulation, and aggregated tau.[74]^38^,[75]^39^,[76]^40 The epidemiology of AD was indicated, in the United States, 7 million people are living with AD. As of 2024, 6.9 million Americans aged 65 and older are assessed to have AD. It affects approximately 1 in 9 people over the age of 65. By 2060, this number could grow to 13.8 million.[77]^41 In the case of AD, 5% of individuals between the ages of 65 and 74, 13.2% of individuals between the ages of 75 and 84, and 33.4% of individual aged 85 and above. There is little understanding of the disease’s cellular and molecular complexity, and at the same time, understanding the immune landscape is also necessary for effective diagnosis and treatment. On the other hand, some immunological parameters have been identified in AD.[78]^42 Immune cells such as dendritic cells (DCs), T cells, and macrophages play essential roles in the disease progression of AD. It has been observed that T cells are involved in AD pathology; T cells are dysfunctional during AD progression. The dysfunction of T cells is a significant adaptive immune response crucial for the progression of AD. Dai and Shen reported the same findings in their research that T cells are associated with inflammatory response, and, simultaneously, they might regulate the inflammatory response through the secretion of pro-inflammatory cytokines.[79]^43 In inflammatory response, the pro-inflammatory cytokine secretion causes neuroinflammation, which is a key factor for AD progression[80]^44 Along with the T cell research, several studies have been performed to understand the involvement of CD4^+ T cells and CD8^+ T cells in the pathogenesis of AD. However, it has been noted that the response against the antigen of CD4^+ T cells and CD8^+ T cells depends on the major histocompatibility complex class II or class I molecule.[81]^45 It might create a further response to the antigen. The number of CD4^+ T cells and CD8^+ T cells are found to be higher in AD patients than in healthy individuals. The augmented number of CD4^+ T cells and CD8^+ T cells exists in the parenchyma tissue of the brain and cerebrospinal fluid (CSF). However, it has been observed that, in the brain parenchyma and CSF, CD8^+ T cells are often larger compared to the CD4^+ T cells in these locations. Additionally, researchers identified changes in subsets of CD4^+ T cells. Here, they found that naive cells have decreased and memory cells have increased.[82]^46 These alterations of the CD4^+ T cell subsets might suggest a persistent antigenic challenge.[83]^47 Other immune cells, such as DCs, are involved in the pathogenesis of AD. DCs, especially those originating from monocyte-derived dendritic cells, might contribute to AD pathogenesis. DCs cause neurodegeneration and brain damage and might interact with amyloid-beta (Aβ) peptides. It has been noted that Aβ1-42 is a significant protein that is associated with AD pathology.[84]^48 When DCs are exposed to Aβ1-42, augmented production of inflammatory molecules causes neuroinflammation; subsequently, AD’s pathogenesis was noted.[85]^49 A special type of macrophage, microglia, is present in the brain, and it acts as the primary immune cell. During AD, when Aβ plaques are created, microglia is accountable for removing Aβ plaques through phagocytosis by either engulfing or destroying. When microglia fails to clear Aβ effectively, accumulating Aβ plaques can cause neuroinflammation. Neuroinflammation is essential in AD advancement.[86]^50^,[87]^51 Single-cell studies evaluate the cells individually to assess their differences and properties and provide insights into cell-to-cell variation and population behavior at a cellular level.[88]^52 They have been applied in different analyses and have proven to be informative. Single-cell studies have been used to understand the biology of aging and have proven to be very informative.[89]^53 These studies have been used to understand COVID-19’s lung cell atlas and etiology.[90]^54 Transcriptional profiling of single-cell studies was used to understand the intercommunication in the heart and cellular diversity.[91]^55 Recently, it has been used to understand the molecular landscape of the placenta.[92]^56 Single-cell transcriptomics can be assessed to understand the brain’s cellular heterogeneity during AD.[93]^36 Similarly, in human brain samples (human prefrontal cortex) of 427 older individuals, Mathys et al. developed the atlas from single-cell transcriptomic analysis to understand the cognitive impairment and varying degrees of AD pathology.[94]^37 However, Alzheimer’s immune landscape must be understood using the single-cell transcriptome. This study has three objectives: first, identification of significantly expressed genes in AD using GenAI; second, analysis of significantly expressed genes in AD using integrative bioinformatics; and third, understanding of the immune landscape uses single-cell analysis. In the first part of our study, we used the three GenAI models to identify the significant genes of ChatGPT-4o and Gemini models. Prompt engineering was applied to develop the question asked in three GenAI models. In the second part of our study, the identified genes of AD were analyzed using integrative bioinformatics. Here, the analyses performed include gene network integration and gene enrichment analysis, pathway enrichment analysis and their networking, genes and their different descriptions, and density plot and bar plot. The third part of our study examined the immune landscape using single-cell analysis. Here, we performed the study in different directions using single-cell analysis, such as categorization of the immune cell types, categorization of T cell types, categorization of other immune cell type, understanding the differentially expressed genes related to immune cell type and their single-cell expression, illustrating the sample correlation and cell-to-cell communication of immune cells, and understanding the clonal frequency and unraveling the clonotype cell proportion. Result First part of our study: Identification of significantly expressed genes in AD using GenAI Identification of significantly expressed genes in AD using GenAI We have identified significatly expressed genes in AD using three GenAI models (ChatGPT-4o, Gemini 2.0 and DeepSeek). The genes have been recorded ([95]Table 1). ChatGPT-4o identified 13 genes, Gemini 2.0 identified 10 genes, and DeepSeek identified 17 genes. We have accumulated and combined genes predicted from the three GenAI models, which indicated that 27 significant genes were associated with AD. Furthuremore, these genes were verified from the litarature ([96]Table S1). Table 1. Three GenAI modes suggested remarkably expressed genes in Alzheimer’s disease ChatGPT-4o Gemini 2.0 Flash DeepSeek Prompt engineering-enabled question asked to GenAI remarkable expressed gene in Alzheimer’s disease remarkable expressed gene in Alzheimer’s disease remarkable expressed gene in Alzheimer’s disease Suggested remarkable expressed genes in Alzheimer’s disease APOE (apolipoprotein E), APP (amyloid precursor protein), BACE1 (beta-secretase 1), PSEN1 and PSEN2 (presenilin 1 and 2), MAPT (microtubule-associated protein tau), TREM2 (triggering receptor expressed on myeloid cells 2), SORL1 (sortilin-related receptor), CD33 (cluster of differentiation 33), CLU (clusterin), GFAP (glial fibrillary acidic protein), NEUROD1 (neuronal differentiation 1), and GSK3B (glycogen synthase kinase 3 beta) APOE (apolipoprotein E), APP (amyloid precursor protein), PSEN1 (presenilin 1) and PSEN2 (presenilin 2), genes involved in neuroinflammation (OLR1, CDK2AP1), and genes related to synaptic function (VSNL1, RTN1, FGF12, and ENC1) APP (amyloid precursor protein), MAPT (microtubule-associated protein tau), APOE (apolipoprotein E), BACE1 (beta-secretase 1), PSEN1 and PSEN2 (presenilin 1 and 2), TREM2 (triggering receptor expressed on myeloid cells 2), SORL1 (sortilin-related receptor 1), INPP5D (inositol polyphosphate-5-phosphatase D), CLU (clusterin), GAB2 (GRB2-associated binding protein 2), HLA-DRB1 (major histocompatibility complex class II, DR beta 1), SNAP25 (synaptosomal-associated protein 25), BDNF (brain-derived neurotrophic factor), and pro-inflammatory cytokines (IL-1B, IL-6, and tumor necrosis factor) [97]Open in a new tab Second part of our study: Analysis of significantly expressed gene in AD using integrative bioinformatics Gene network integration and pathway enrichment Gene network integration was developed using 27 identified genes and GenAI combined genes. We developed two models, direct interactions analysis ([98]Figure 1A) and shared neighbors ([99]Figure 1B). Furthermore, we developed pathway enrichment analysis shows two models, force-directed layout ([100]Figure 1C) and graphopt layout ([101]Figure 1D). During pathway enrichment analysis, we found several overlaps. To understand the overlap, we used a regression plot using XD-score vs. significance of overlap (Fisher test, q value) ([102]Figure 1E). Here, we found that the absolute Pearson correlation between XD-scores and Fisher q values is 0.9 and XD-score significance threshold is 0.37. Figure 1. [103]Figure 1 [104]Open in a new tab Gene network integration pathway enrichment analysis and gene enrichment analysis using 27 identified and combined genes from three GenAI models (A) Network integration depicts direct interactions. (B) Network integration depicts shared neighbors. (C) Pathway enrichment analysis depicts force-directed layout. (D) Pathway enrichment analysis depicts graphopt layout. (E) Regression plot to understand the overlap using XD-score vs. significance of overlap. (F) Pathway enrichment analysis associated with different pathways along with Alzheimer’s pathway with the highest −log10 (FDR) value (value = 6). (G) Pathway enrichment analysis depicted networking between the pathways shows that Alzheimer’s pathway is the main hub. Pathway enrichment analysis and their networking Pathway enrichment with fold enrichment analysis was performed using the list of genes, and the outcome has been noted ([105]Figure 1F). This study shows that the listed genes are associated with different pathways: Notch signaling pathways, interleukin (IL)-17 signaling pathways, etc. The analysis shows that the genes are associated with Alzheimer’s pathway with the highest −log10 (false discovery rate [FDR]) value (value = 6). At the same time, the networking between the pathways was depicted using pathway enrichment analysis that shows that Alzheimer’s pathway is the main hub in the network ([106]Figure 1G). Genes and their different descriptions, density plot, and bar plot We developed a list of the genes and their general descriptions, such as Ensembl Gene ID, Entrez, associated chromosome, position in the chromosome, description, etc. ([107]Table S2), and different descriptions at genetic and molecular levels, such as percentages of GC content, transcript count, genome span, CDS (coding sequence) length, transcript length, 5′ UTR, 3′ UTR, and exon (nExons) ([108]Table S3). We also developed density plots of the genes, where genes are compared with the rest of the genome. The density plots include coding sequence length ([109]Figure 2A), transcript length ([110]Figure 2B), genome span ([111]Figure 2C), 5′ UTR length ([112]Figure 2D) and 3′ UTR length ([113]Figure 2E), and percentages of GC content ([114]Figure 2F). Figure 2. [115]Figure 2 [116]Open in a new tab Different density plots and bar plots were developed using 27 identified and combined genes from three GenAI models, where genes are compared with the rest of the genome (A) Density plot shows coding sequence length. (B) Density plot shows transcript length. (C) Density plot shows genome span. (D) Density plot shows 5′ UTR length. (E) Density plot show 3′ UTR length. (F) Density plot show percentages of GC content. (G) Bar plot show the number of exons. (H) Bar plot show the number of transcripts per gene. Again, we developed some bar plots, such as the number of exons (chi-square tests, p = 0.74) ([117]Figure 2G) and the number of transcripts per gene (chi-square tests, p = 0.14) ([118]Figure 2H). Both plots were developed using chi-square tests. Third part of our study: Understanding of immune landscape using single-cell analysis Overall categorization of the immune cell types Here, we found the distribution of the percentage of cells such as B cells (1.49%), effector CD8^+ T cells (33.42%), effector memory T cells (4.30%), naive T cells (45.95%), macrophages (11.60%), and DCs (3.24%) ([119]Figure 3A). Figure 3. [120]Figure 3 [121]Open in a new tab Overall categorization of the immune cell types using single-cell analysis (A) The percentage of the immune cells. (B) Overall categorization of the single-cell type. (C) Overall healthy immune cell types. (D) Overall immune cell types in Alzheimer’s disease. Now, we compare the overall categorization of the single-cell type ([122]Figure 3B), overall healthy cell ([123]Figure 3C), and cell from AD ([124]Figure 3D) and find the clustering of B cell, CD8^+ T cell, effector memory T cells, naive T cell, macrophages, and DCs. Based on clustering, we found two cell groups with the highest clustering: naive T cell and CD8^+ T cell. Categorization of the T cell type populations Again, we have categorized the three T cell populations, i.e., effector CD8^+ T cell, effector memory T cell, and naive T cell. First, we observed effector CD8^+ T cell, both healthy effector CD8^+ T cell population ([125]Figure 4A) and AD’s CD8^+ T cell population ([126]Figure 4B). We found that the number of effector memory T cell population is less in AD patients compared to healthy individuals. Similarly, we found effector CD8^+ T cells, both healthy effector memory T cell population ([127]Figure 4C) and Alzheimer’s effector memory T cell population ([128]Figure 4D). Here, we found two cluster effector memory T cell populations. During the comparison of both populations, we found that the number of effector memory T cell populations is lower in AD patients compared to healthy individuals. Figure 4. [129]Figure 4 [130]Open in a new tab Categorized the three T cell type populations and other cell type populations using single-cell analysis (A) Healthy cell effector CD8^+ T cell population. (B) Alzheimer’s disease CD8^+ T cell population. (C) Healthy cell effector memory T cell population. (D) Alzheimer’s effector memory T cell. (E) Healthy cell naive T cell population. (F) Alzheimer’s naive T cell population. (G) The population of B cells in healthy individuals. (H) The population of B cells in Alzheimer’s disease individuals. (I) The population of macrophages in healthy individuals. (J) The population of macrophages Alzheimer’s disease. (K) The population of dendritic cells in healthy individuals. (L) The population of dendritic cells in Alzheimer’s disease individuals. Again, we found naive T cells, both the healthy naive T cell population ([131]Figure 4E) and Alzheimer’s naive T cell population ([132]Figure 4F). Here, we found two clusters of naive T cells. During the comparison of both populations, we found the same thing: the number of effector memory T cells has gone down in the case of AD patients compared to healthy individuals. However, it is clear that, among these three cell categories (effector CD8^+ T cell, effector memory T cell, and naive T cell), effector memory T cell has two subsets of cells. Categorization of other immune cell type populations Here, we categorize B cells, macrophages, and DCs. We noted the population of B cells in healthy individuals ([133]Figure 4G) and B cells in AD individuals ([134]Figure 4H). Similarly, we noted the population of macrophages in healthy individuals ([135]Figure 4I) and the population of macrophages in AD individuals ([136]Figure 4J). Again, we noted the population of DCs in healthy individuals ([137]Figure 4K) and the population of DCs in AD individuals ([138]Figure 4L). Here, we found that the B cell population has decreased in Alzheimer’s patients compared to healthy individuals. The macrophage population has started to spread in Alzheimer’s patients compared to healthy individuals. Finally, we noted that DCs have also decreased in Alzheimer’s patients compared to healthy individuals. Understanding the differentially expressed genes related to immune cell type and their single-cell gene expression study Again, we noted the top ten differentially expressed genes related to immune cell type concerning the top ten p values ([139]Table S4). Again, we depicted the single-cell expression of those genes, and they are NDUFV2 ([140]Figure 5A), CAT ([141]Figure 5B), MRPS34 ([142]Figure 5C), PBX3 ([143]Figure 5D), THOC2 ([144]Figure 5E), CCDC57 ([145]Figure 5F), PBXIP1 ([146]Figure 5G), SDHAF3 ([147]Figure 5H), PPP4C (([148]Figure 5I), and MAP3K8 ([149]Figure 5J). It shows the variation of expression of ten genes. Figure 5. [150]Figure 5 [151]Open in a new tab Differentially expressed genes (DEG) associated with immune cell type and their single-cell expression were evaluated through the single-cell expression study. It also illustrates the sample correlation and cell-cell communications of immune cells (A) DEG shows the expression of NDUFV2 (B) DEG indicates the expression of CAT (C) DEG shows the expression of MRPS34 (D) DEG demonstrates the expression of PBX3 (E) DEG indicates the expression of THOC2 (F) DEG shows the expression of CCDC57 (G) DEG indicates the expression of PBXIP1 (H) DEG illustrates the expression of SDHAF3 (I) DEG demonstrates the expression of PPP4C (J) DEG indicates the expression of MAP3K8. (K) The single-cell expression study shows the sample correlation between immunological cells represented through the heat map (L). The single-cell expression study depicted a cell-to-cell communication plot in immune cells of Alzheimer's disease considering the width (M) depicted cell-to-cell communication plot in immune cells of Alzheimer's disease considering the count. Study to illustrate the sample correlation and cell-to-cell communications of immune cells We have depicted the sample correlation of immune cells represented through the heatmap ([152]Figure 5K). The study understood intercellular signaling, or cell-to-cell communication, which indicates how cell communication occurs. Cells communicate through various mechanisms, such as chemical signaling or straightforward contact.[153]^57 Cell-to-cell communication transmits their activities and maintains the processes of tissues and organs. It may also use extracellular vesicles. We noted the cell-to-cell communications in the immune cells of AD, their ligands, and receptors concerning the top ten p values ([154]Table S5). Furthermore, we depicted the cell-to-cell communication plot in immune cells of AD considering the width ([155]Figure 5L) and count ([156]Figure 5M). Understanding the clonal frequency Clonal frequency refers to the proportion of a specific cell population (or clone) within a bigger group of cells.[157]^58 Within a sample, clonal frequency indicate the proportion of various cell populations (clones), providing a population’s clonal diversity, and revealing its evolutionary history.[158]^59^,[159]^60 We analyzed the clonal frequency of immune cells of AD, the overall clonal frequency of the healthy population ([160]Figure 6A), and the overall clonal frequency of the Alzheimer’s population ([161]Figure 6B). We developed the clonal frequency of the healthy CD8^+ T cell population ([162]Figure 6C) and Alzheimer’s CD8^+ T cell population ([163]Figure 6D). Again, we depicted the clonal frequency of the healthy naive T cell population ([164]Figure 6E) and the clonal frequency of Alzheimer naive T cell ([165]Figure 6F). Similarly, we depicted the clonal frequency of healthy effector cell memory T cell population ([166]Figure 6G) and the clonal frequency of Alzheimer effector memory T cell ([167]Figure 6H). Here, the clonal frequency of the effector memory T cell population is less than that of other populations. We depicted the clonal frequency of macrophages in healthy individuals ([168]Figure 6I) and in individuals with AD ([169]Figure 6J). Similarly, we depicted the clonal frequency of the population of DCs in healthy individuals ([170]Figure 6K) and the clonal frequency of DCs in individuals with AD ([171]Figure 6L). Here we found almost a low or no clonal population of DC. Finally, we depicted the clonal frequency of the population of B cells in healthy individuals ([172]Figure 6M) and the clonal frequency of B cells in Alzheimer’s individuals ([173]Figure 6N). Here, we have also not found the clonal population. The prominent clonal frequency was found in two groups of cell populations, which are both CD8^+ T cell populations (healthy CD8^+ T cell population and Alzheimer’s CD8^+ T cell population) and naive T cell population (healthy naive T cell population and the clonal frequency of Alzheimer naive T cell). Figure 6. [174]Figure 6 [175]Open in a new tab Immune cell-associated single-cell study of AD illustrates the clonal frequency of a distinct immune cell population and the clonotype cell proportion (A) the overall clonal frequency of the healthy population, (B) the overall clonal frequency of the Alzheimer population, (C) the clonal frequency of healthy CD8+T cell population, (D) the clonal frequency of Alzheimer CD8+T cell population (E) the clonal frequency of healthy naive T-cell population (F) the clonal frequency of Alzheimer naive T cell (G) the clonal frequency of healthy effector cell memory T cell population (H) the clonal frequency of Alzheimer effector memory T cell (I) the clonal frequency of macrophages in healthy individuals (J) the clonal frequency of macrophages in Alzheimer's individuals (K) the clonal frequency of population of dendritic cells in healthy individuals (L) the the clonal frequency of dendritic cells in Alzheimer's individuals (M) the clonal frequency of B cell in healthy individuals (N)the clonal frequency of B cell Alzheimer's individuals. (O) the clonotype cell proportion from one healthy individual, (P) the clonotype cell proportion from another healthy individual, (Q) the clonotype cell proportion from one Alzheimer's individual, (R) the clonotype cell proportion from other Alzheimer's individual. Unraveling the clonotype cell proportion Clonotype cell proportion shows the percentage or fraction of cells in a sample that belongs to a particular clonotype. It measures how much a particular cell clone represents the overall cell population.[176]^61^,[177]^62 We analyzed the clonotype cell proportion from two healthy individuals’ data ([178]Figures 6O and 6P). Similarly, we have analyzed the clonotype cell proportion data from two AD individuals ([179]Figures 6Q and 6R). In both groups of clonotype cell proportion, only clonal cells are found in the CD8^+ T cell population (healthy CD8^+ T cell population and Alzheimer’s CD8^+ T cell population) and naive T cell population (healthy naive T cell population and the clonal Alzheimer’s T cell population). Discussion Novelty of our study In this work, we present our remarkable achievement that the GenAI models can successfully fetch the significant genes of AD. We used the following three GenAI models to successfully fetch the 27 significant genes of AD: ChatGPT-4o, Gemini model, and DeepSeek. ChatGPT-4o determined 13 genes, Gemini 2.0 determined 10 genes, and DeepSeek determined 17 genes. In addition, we revealed that ChatGPT-4o and DeepSeek identified several common genes. Most genes identified in GenAI models are involved in the progression and perturbation of a disease. These 27 genes were further analyzed using integrative bioinformatics such as gene network integration and gene enrichment analysis, pathway enrichment analysis and their networking, and genes and their different descriptions such as molecular biology or genetic characteristics. Another part of the study has found the immune landscape of AD using single-cell analysis. Here, we performed the study in different directions using the single-cell analysis, such as categorization of the immune cell types, categorization of the T cell types, categorization of another immune cell type, understanding the differentially expressed genes related to immune cell type and their single-cell expression, study to illustrate the sample correlation and cell-to-cell communications of immune cells, and understanding the clonal frequency and unraveling the clonotype cell proportion. Our empirical investigation revealed the immune landscape of AD using this single-cell analysis. Integrating AI models into bioinformatics has recently become a regular practice and brought about a revolutionary era.[180]^63 Researchers amalgamate AI and bioinformatics techniques to fulfill the research objective. Researchers have also combined AI and bioinformatics techniques to solve critical research objectives. In the breast cancer landscape, Wager et al. used AI and bioinformatics techniques to map the cyclin-dependent kinase 4/6 inhibitor biomarker.[181]^64 We also used prompt engineering associated with LLM or MLLM and integrative bioinformatics techniques to understand significant antibody escape mutations in SARS-CoV-2. Here, we applied the MLLM or LLM to demonstrate the important antibody escape mutations in NTD and RBD of the S-protein of the SARS-CoV-2. Here, we found 15 NTD significant mutations and 17 RBD point mutations and characterized the mutations using bioinformatics. Furthermore, the study characterized those mutations in terms of count, distribution, ΔΔG of mutation (ΔΔGstabilitySDM, ΔΔGstability DUET, and ΔΔG stability mCSM) to assess the destabilization or stabilization of mutation, distance to PPI (protein protein interaction) interface, interaction interface, change in the flexibility, and Δ vibrational entropy energy using bioinformatics. However, the study is an excellent example of the fusion of AI and bioinformatics techniques to fulfill the research objectives.[182]^5 In this study, we amalgamate AI models into bioinformatics to understand the genetic landscape. Furthermore, we also used single-cell techniques to understand the immune landscape of AD. Therefore, this study is highly novel and significant. Findings from this study and their importance This study has tried to identify the immune landscape of AD using single-cell analysis. The study demonstrated a high percentage of effector CD8^+ T cells (33.42%) and naive T cells (45.95%) compared to other populations. It noted that effector memory T cells have two subsets of cells among the T cell population and the macrophage population has started to spread in Alzheimer’s patients compared to healthy individuals. At the same time, it also indicated that DCs have been decreased in Alzheimer’s patients compared to healthy individuals. The differentially expressed genes related to immune cell type and their single-cell gene expression study reveal that the top ten highly expressed genes with the highest p value are NDUFV2, CAT, MRPS34, PBX3, THOC2, CCDC57, PBXIP1, SDHAF3, PPP4C, and MAP3K8. The clonal frequency indicates that both healthy and Alzheimer’s CD8^+ T cell populations, as well as healthy and Alzheimer’s naive T cell populations, show the highest clonal frequency. This was further observed in the clonotype cell proportion study. Mathys et al. have performed a single-cell transcriptomic study to illustrate a blueprint to understand more about AD’s cellular and molecular basis. They provided a detailed view of the cellular diversity of older people’s prefrontal cortex. The study also understood the cellular subpopulations.[183]^36 Another study by Mathys et al. used the single-cell technique to understand AD pathology. The study has successfully generated the single-cell transcriptomic atlas of 2.3 million prefrontal cortex cells of aged humans. They also identified the inhibitory neuron subtypes depleted in AD. They show the link between inhibitory neurons and resilience to AD pathology.[184]^37 Similarly, Brown et al. performed another study on single-cell levels in AD, using nearly 74,000 cells of human cortex samples (cortex prefrontal and cortex entorhinal samples). It shows the AD-related transcriptomic changes.[185]^65 Again, Mathys et al. analyzed aged brains containing 1.3 million cells from 283 brain samples of postmortem human covering 48 individuals in another study. They have illustrated the vulnerability of these neurons in AD. They indicated that the reelin signaling pathway was associated with the susceptibility of these neurons. Here, they have identified an astrocyte linked to the cognitive function to AD pathology.[186]^66 However, our study illustrated the immune landscape of AD using the single-cell analysis. Together with the immune cell population, we studied cell-to-cell communication, and the top ten highly expressed genes considering the highest p value using the single-cell analysis. Therefore, our single-cell study is vital to describe the immune landscape of AD. Conclusion Our study illustrated Alzheimer’s genetic and immune landscape using modern technology like GenAI, integrative bioinformatics, and single-cell analysis. To understand Alzheimer’s pathological landscape, we used the amalgamate of GenAI, bioinformatics, and single-cell analysis techniques. Our study’s main question was as follows: Did GenAI fetch significant AD genes? GenAI can retrieve the remarkable genes that help in the disease progression of AD. However, our novel strategy using the amalgamation of GenAI, bioinformatics, and single-cell analysis can help the genetic and immune landscapes of AD. On the other hand, our single-cell analysis highlights the population of immune cells and their interaction. They also describe the subset of some immune cell populations. At the same time, our approach delivers a solid foundation for a study of the immunology basis of the disease. The study also found heterogeneity within the immune cell population. We provide new strategies to understand Alzheimer’s genetic and immune landscape, using modern technologies like GenAI and single-cell profiling, which can highlight the complexity of AD pathology. At the same time, the new strategies assisted in the quick understanding of the disease landscape of AD. Following our strategy, future studies can be performed to understand the genetic and immune landscape of other diseases by quickly identifying genes and their immune landscape associated with disease progression, which will also help further in next-generation marker development and therapeutic discovery. Materials and methods Overview of the study The study has three parts. First, we have identified the remarkable expressed genes in AD using the prompt engineering-enabled GenAI. Second, the identified genes have been analyzed using integrative bioinformatics. Third, the immune landscape of AD has been analyzed using single-cell analysis. First part of our study: Identification of significantly expressed genes in AD using GenAI Significantly expressed genes in AD using the prompt engineering-enabled GenAI A considerable number of genes have been expressed in AD. It is tough to understand the significant genes. We used different GenAI models to understand the remarkable expressed gene in AD. Our central question in this study was as follows: Can prompt engineering enable GenAI to identify expressed genes in AD? To answer this question in this section of the study, we have used three GenAI models, which are OpenAI's ChatGPT (ChatGPT-4o),[187]^10^,[188]^67 Google DeepMind Gemini model,[189]^17 and DeepSeek.[190]^68 Here, we have used prompt engineering-enabled questions to the GenAI. Prompt engineering Prompt engineering, a crucial method in NLP, is an area of computer science, especially in the research of the AI domain. It helps create prompts to use GenAI effectively. Prompt engineering optimizes and increases the performance of GenAI models. Previously, we generated proper questions for GenAI using prompt engineering, which helped us generate very straightforward answers from GenAI.[191]^5^,[192]^6^,[193]^69^,[194]^70^,[195]^71^,[196]^72 In most circumstances, we tried to utilize few-shot prompts during our question-framing.[197]^72^,[198]^73 Second part of our study: Analysis of significantly expressed gene in AD using integrative bioinformatics In the second part of our study, we analyzed significantly expressed genes in AD using integrative bioinformatics. Here, we analyzed the accumulated genes reported by the three GenAI models. Gene network integration and gene enrichment analysis We developed a gene network using the GeneMANIA prediction server, which used the gene significantly expressed in AD. It was collected from organism-specific functional genomics datasets, BioGRID, GEO, Pathway Commons, IRefIndex, and I2D. It was used for gene function prediction. Using gene query, it developed the prioritizing genes gene network for functional assays.[199]^74^,[200]^75 Gene enrichment analysis was done using EnrichNet, a network-based enrichment analysis.[201]^76 Pathway enrichment analysis and their networking Using the list of the genes, pathway enrichment with fold enrichment analysis was performed using the ShinyGO server.[202]^77 Pathway networking was developed using the same server. Genes and their different descriptions, density plot, and bar plot We have developed the list of the genes, their general descriptions, and genetic and molecular levels descriptions, and analysis was performed using the ShinyGO server.[203]^77 Again, the characteristics of genes were compared with the rest in the genome, and the density plots were developed. The third part of our study: Understanding of the immune landscape using single-cell analysis Single-cell dataset Single-cell analysis was performed using the CSF of human samples from healthy individuals and interindividuals with AD. Gate et al. collated the data of 164 human subjects from three different cohorts to understand the adaptive immune response during the pathogenesis of AD[204]^78 and submitted the data to the GEO database ([205]GSE134577 ). The data were acquired using the Illumina NovaSeq 6000 (Homo sapiens) platform ([206]GPL24676). We used the aforementioned dataset (Illumina NovaSeq 6000) for further single-cell RNA sequencing analysis. Single-cell analysis A single-cell study was performed using scImmOmics, which can acquire and curate scRNA-seq (single-cell RNA sequencing) datasets. The server records over 2.9 million cell type-tagged immune cells fetched from seven single-cell sequencing technologies. This is a comprehensive platform to assess the diversity and heterogeneity of immune cells.[207]^79 The single-cell analysis study was conducted from different directions, such as categorization of the immune cell types, categorization of the T cell types, categorization of another immune cell type, understanding the differentially expressed genes related to immune cell type and their single-cell expression, illustrating the sample correlation and cell-to-cell communication of immune cells, understanding the clonal frequency, and unraveling the clonotype cell proportion. A flowchart was depicted to illustrate the comprehensive methodology of our analysis ([208]Figure 7). Figure 7. [209]Figure 7 [210]Open in a new tab A flowchart illustrating the comprehensive methodology of our analysis Data availability All data are included within the manuscript. Acknowledgments