Clinical Genomic Analysis and Diagnosis --Genomic Analysis Ex Vivo, in Vitro and in Silico

Seven years ago, I systemically reviewed single cell techniques with genomic and proteomic analyses which was called Single-Cell Genomic Analysis. After many years of arduous work, single cell techniques with downstream genomic and proteomic analysis have been applied to clinical fields including molecular pathology, molecular genetics, forensic medicine and biomarker d iscovery. On top of that, dynamic cell-sorting technique combined with downstream cell culture and genomic analysis of stem cell for regeneration medicine and cancer stem cell fo r d ifferentiation have also been greatly developed in clin ical fields. More importantly, tissue level sampling with in silico analysis has been applied in therapeutic targeting for advanced neoplastic disease. Recent development in sorting homogeneous cells in vitro (or single cells technique), ex vivo (dynamic analysis or small number of cell cu lture with downstream genomic analysis) and insilico (tissue level sampling with in silico analysis) have allowed physician scientists with a choice to select one of these above techniques with genomic analysis to apply to their clin ical research fields. To fully understand these modern techniques, this manual will review recently developed methods or clinical genomic analysis in vitro, in silico and ex vivo. In the review paper, I will also introduce how to utilize these techniques in different clinical fields. The manual will also address some of the challenges for clinical genomics analysis and diagnosis due to mixed cells from clinical specimens.


Introduction
The mixed cell population in clinical samp les can mask the actual results of genetic diagnosis and genomics analysis. In order to overcome the challenges in purity and limited cell counts after in it ial purification , s ingle cell technique, or similar techniques, have now being widely developed for over a decade. One of the most common is single cell PCR [1]. Due to advancements in techniques in amplify ing genomic DNA/ RNA and signal magnify ing fo r protein fro m s mall number of cells, these techniques, including genomics and proteomics are now widely being emp loyed in the clinical fields [2]. Based on research and development for genomic analysis from clinical samp les, physicians and scientists are studying faster/p urer t echn iques, such as laser cap tu re microscopy, along with downstream geno mic analysis to study clinical specimens. The single-cell harvest obtained fro m glass slides combined with downstream genomics and proteomics have been developed in mo lecu lar pathology, mo lecular genetics and forensic medicine [3]. So me call the technique, cytogenomics [4]. The technique is named as in vitro analysis because the procedure defined as in vitro (Latin: within the glass or glass slide) is performed outside the living organis m. On the other hand, dynamically pure cell-sorting technique combined with ex vivoculture with downstream genomic analysis (such as stem cell for regeneration medicine and cancer stem cell for inducing differentiation) has also been tremendously developed [5]. This technique is known as ex vivo because the measurements are done in an artificial growth environ ment (outside the organism) with minimu m alterations of the natural conditions. Moreover, after Dr. Sch mid first used "micro-d issection in silico" to analyze gene expression profiles to uncover bio markers fro m clin ical specimens [6], tissue level samp ling by in silico analysis is now increasingly being applied in genomic analysis of heterogeneous cells fro m clin ical t issues [7]. Here the technique is called as in silico genomic analysis due to the performance via computer or program simu lation, as an analogy to the Latin phrases in vivo and in vitro refer to experiments done in living organisms and outside of living organisms, respectively. In order to incorporate all the clinical genomic techniques by using these three methods, in vitro, ex vivo and in silico, I will refer to clinical genomic analysis as genomic analysis and diagnosis by in vitro, ex vivo and in silico. Geno mic analysis in vitro, used in molecular pathology and molecular genetics from pretreated tissue [8], co mbines cell isolation and genomic analysis fro m clin ical specimen. Geno mic analysis ex vivo, encompass three processes, 1) liv ing cell separation, 2) culture and 3) downstream genomics analysis as well. Now, FACS/MACS for living cell separation can be combined with fixed cell immunohistochemical staining (IHC) [9] and laser micro -dissection technique (orig inally only used for fixed cell) can be utilized in liv ing cell with downstream cell culture [10]. Geno mic analysis in silico, direct tissue-level specimens for genomic analysis, in which several modules in b ioinformatics are able to identify specific genomic pro file in the mixed tissue level, Here, in order to better understand clinical genomic analysis and diagnosis fro m clinical specimens, I will co mpare genomic analysis among intro, ex vivo and in silico for different clin ical specimensand different purposes. Based on their conditions and purposes, clinical scientists can decide which way is the best genomic analysis for their specimens.

Clinical Genomic Analysis by intro, ex vivo, and in Silico
In my previous review paper [11], it was shown that the flow-cytometric cell sorting, magnetic cell separation (FACS/MACS) and laser-based micro-dissection of tissues provide the basic methods to isolate similar cells to study gene expression profiling fro m clin ical specimens. FACS isolating technique for cells in solution labelled with fluorescent signals can sort these cells with a specific biomarker such as CD3/CD4/CD8 for ly mphocytes and CD133/CD34 for stem cells and cancer stem cells (CSCs). At present, multi-colored fluorescence-activated cell sorters (mult i-coloured FACS) can select ively separate and collect homogeneous cells with identical phenotypic features in a collection tube so that FACS can increase its ability to study gene expression profiling in a g iven cell type [12]. In the other fields, laser micro-d issection technique to isolate cells on glass slides labelled with fluorescent signals or markers can sort clinical cells that rely on specific mRNA/protein biomarkers and morphology change such as tumor cells or cancer stem cells. The characteristics of laser micro -dissection have allowed us to quickly study a given cell in vivo localization and to analyse the cell's microenvironment in vivo [13].
Relied on research and development, the techniques for sorting single-cell or a s mall number o f homogeneous cells fro m clin ical specimens have been developed and categorized by the methods, purposes and functions described in Table 1.
Here I will introduce each method,in details, along with this advantage and disadvantage in the following order: (1) genomic analysis ex vivo covering cell-sorting with downstream culture ex vivo and genomic analysis; (2) genomic analysis in vitro including treated cell (fixed and stained cells) and un-treated cell and (3) direct genomic analysis in silico for clin ical specimens. *In vitro means the technique of performing a given procedure in a glass or glass slides environment outside of a living organism; in vivo is experimentation using a whole or living organism; ex vivo is experiment done for tissue in an arti ficial environment outside the organism; in silico means an expression or experiment "performed on computer or via computer simulation".

Genomic Analysisex vivo
Cell culture fro m clin ical specimens is different fro m cell lines. Cells that are cultured directly fro m clinical specimens are known as primary cells. The primary cells isolated fro m tissues can be cultured ex vivo or in vivo environ ments. The primary cell cultures can be often used in ly mphocytes, cancer stem cell, stem cells and primary cancer cells [14].
Along with the research of stem cells for regenerative med icine, the study of cancer stem cells for bio marker discovery and the application of T-cell for passive and active immunotherapy as shown in Table 2, stem cells, cancer stem cells, and T-cells harvested from clinical specimens with genomic analysis can play an increasingly important role in biological and medical fields. Although ex vivo differentiation of stem cells intospecial somatic cell types such as neurons and myocytes has a well-established model, long-term cultureof ex vivo autogenously adultstem cells so far have not been co mpletely successful. The extremely low number of these cells in p rimary hematopoietic organs and the lack of good culture systems that support proliferat ion of undifferentiated stem cell have influenced their b iological research and clinical application. In 1999, we reported cancer cell cultured in high-dose radiated mice to increase the cell number to study the cell characteristics [15]. Encouragingly, now repopulating capacityof ex vivo culture of hematopoietic stem cells has also largely been reported, which will indicate a good future for regeneration med icine [16]. In addition, ex vivo culture and long-term storage of induced pluripotentstem cells(IPS cells) have been tremendously reported.

Genomic Analysis in vitro
Since 1976, laser micro-dissection has been increasingly applied in different fields as shown in Table 3. Currently, three microdissection methods have been routinely emp loyed. Those are (1) laser-assisted mechanical t issue micro-dissection [17], (2) laser p ressure catapult micro -disse ction [18] and (3) laser capture micro-dissection [19]. Most laser-based microdissections require tissue pretreatment such as fixation, dehydration and staining. The application of these techniques requires fully consideration to downstream work such as DNA array CGH, mRNA microarray and proteomics. The advantage of laser-based microdissections can be comb ined with Ab-based immunohistochemical/imm unocytochemical (IHC/ICC) or DNA FISH and RNA FISH staining to increase the cell specificity fro m their bio marker staining. In these several years, following develop mental requirement of cancer stem cells for bio marker identification and therapeutic targeting and the applicat ion of stem cells for regeneration med icine, laser-based micro-d issection will be faced with two large developments: (1) as indicated above, process from cultured liv ing cells, which can continue culture for downstream genomic analysis; (2) co mb ined with automation system, whose process can be used for high-throughput screening such as for bio marker discovery [20].
Cell-sorting technique by FACS and MACS is also quickly developed. With FA CS technique development of mu lti-colored fluorescence-activated cell sorters and increasing antibody products, the FACS technique can play a much mo re important role in cell-sorting technique for genomics and proteomics analysis of clinical specimens [21]. Nowadays, two fields are being quickly developed, the one is based on labelling mRNA inside cells to sort cells by florescent-labelled mRNA and the other one is combined with automation system, the purpose is high throughput screening to discover new bio markers [22].
Due to impurities of FACS and MACS in clin ical specimens, we have added a single cell man ipulation to increase the purity after MACS or FACS process so downstream-cells can provide mo re accurate data in genomic signature analysis and diagnosis for patient's therapeutic targeting [23].

Genomic Analysis in silico
In the clinical field, clin ical specimens are often and directly frozen due to requirements of pathological processes. If the specimens are processed at the tissue level for microarray by total mRNA, array CGH by total genome DNA and proteomics by total protein, genomic analysis in silico is an exclusive way for the analysis of genomic data because the genomic data at the tissue level are mixed with genome fro m different cells [24]. According to our experiences, at least three ways can be used for genomic analysis for heterogeneous cells: h ierarch ical cluster, principle co mponent analysis (PCA), and self-organizing map (SOM). Hierarchical clustering can be quickly processed due to its usage of similar exp ression patterns for cell groups. Princip le co mponent analysis (PCA) is primarily aimed at finding co mp lex relationships between variables in a dataset so it has been extensively applied in bio marker discovery of known cells fro m mixed-cells tissues. This feature helps us to study variables and factors that are not correlated to each other [25]. The self-organizing map (SOM) is a powerful tool for grouping and visualizing high-dimensional co mplex data which can be applied for a two-dimensional plane [26]. Fo llo wing ext racting data fro m mixed cells of clinical specimens, we have routinely emp loyed two of the three platforms to analyse data from the mixed cells as shown in Table 4. If genomic data fro m the clin ical specimens do not contain a series of control samples to process micro-d issection in silico, we can use clinical bioinformat ic module co mbined with bio logical or medical biomarkers corresponding this disease to analyse the data, such as using "supervised leaning" protocol.  We have studied a set of geno mic data (genomic analysis in silico via genomic analysis in vitro, or single cell genomic analysis) fro m similar samp le as Figure 1, both will produce accurate intersection results. The results show that genomic analysis in silico can mine three digit genes. This method is highly sensitive that this can be used to uncover genome profiles; on the other hand, single cell genomic analysis can detect genes, which is so specific that this can be utilized to confirm the genomic p rofile.

Regeneration Medicine
A major issue in regenerative medicine is the cell sources used to rebuild damaged t issues. Although many questions are still unanswered, most physicians and scientists nowadays have reported that regeneration in humans from the body's own tissues is feasible. Regeneration is a regulative developmental process ubiquitous across all human organs. It functions throughout the life cycle to restore the normal function of cells, tissues, organs, appendages and whole organisms.The regenerative capability is absent or low in some cells such as hepatocytes, myofibers, osteocytes, and most neurons [27]. We know that three sources of cells, induced pluripotentstem cells(IPS cells), emb ryonic stem cells (ESCs) and adult stem cells (ASCs) can be used for cells and/or organs regeneration for damaged tissues. ESCs are in vitro cultivated pluripotent cells derived fro m the inner cell mass (ICM) of the embryonic b lastocyst. ES cells can be differentiated into representative derivatives of all three embryonic germ layers (endoderm, ectoderm and mesoderm) in both in vitro and in vivo [28].Adult stem cells also can mu ltiply to regenerate the definit ive cells to replace damaged tissues. They have been found in normal scaring after injury, such as in myocardiu m, pancreatic islet cells, spinal cord and retina t issues [29]. Because stem cells (ESC and ASC) have high growth and self-renewal capacity, several papers and protocols have been successfully reported to use the cells into the therapeutic possibilities, such as diabetes, mycard iac infarct ion and Parkinson's disease.
Genomic analysis ex vivo can play a critical role in proliferation and differentiation of stem cells for regenerative med icine, such as dynamic mon itoring genomic profile for a s mall amount of cultured cells. After we understand the genome and the corresponding pathway from the primary cells in a differentiating period, we can select corresponding growth factor or compound to control cell proliferation and to target differentiat ion based on the genomic analysis.

Sing le Cell Diagnosis Via Single Cell Geno mic Analysis/Diagnosis
The pathology diagnosis relies on cell mo rphology and cell arrangement such as tumor cell diagnosis based on the cell morphological change and infiltrat ing into normal tissue. The diagnosis of molecular genetics and cytogenetics prefers chromosome structure for analysis. Following the development of single cell techniques, a new term, "Single Cell Diagnosis", has emerged in mo lecular pathology and mo lecular genetics/cytogenetics [30]. As shown in Table 5, single cell diagnosis can be utilized for several fields, such as pathology, genetics and microbiology. Now, this technique can be used in many clinical laboratories, for example, surgical specimens fro m operation, biopsy specimen fro m internal medicine (especially in haematology), blood samples fro m paed iatrics (perinatal diagnosis), obstetrics/gy naecology and psychiatrics as well [31]. Single cell genomic analysis and diagnosis extended fro m single cell technique have some advantages over single cell diagnosis, for instances, (1) fo llo wing genomic analysis fro m single cell harvest, it can convert pathological changes into biomarker discovery for genetic and tumor disease; (2) along with genomic analysis, single cell geno mic analysis and diagnosis can lin kpathogenesis of disease and therapeutics so single cell geno mic analysis and diagnosis can be used for personalized med icine; (3) single cell genomic analysis and diagnosis can extend beyond genetic disorder and cancer disease to diagnosis of several other diseases.

Forensic DNA analysis via Single Cell DNA-profiling
The characteristics and the inheritance of autosomal and sex chro mosomes are the basis to determine DNA typing in forensic medicine. Forensic DNA analysis, such as, restriction frag ment length poly morphisms (RFLP), short tandem repeats (STR), variab le number tandem repeats (VNTR), and mitochondrial DNA has been routinely applied into forensic laboratory. Along with human genome decoded and development of forensic-DNA techniques, the application of forensic-DNA analysis have extended from a blood sample into DNA trace specimen such as single cells or small nu mber of cells. In RFLP technology, the polymorphis m can be read by HPLC and DNA-Sequencer in a specific gene or non-coding DNA sequence in a indiv idual person [32]; STR analysis comb ined with PCR technology with fluorescent labels, can automatically detect at a single cell level and analy zed the DNA genomic profile for a great amount of specimens [33]; because VNTR are inherited fro m the individual's parents,, it also can be used for a single cell DNA-pro filing [34]; Since mitochondrial DNA is passed fro m one generation to the next solely through the maternal line of a family, By co mparing the mitochondrial DNA (mDNA)fro m samples of maternal families and maternal relatives, forensic scientists are able to search and demonstrate a specimen trace in a maternal family [35,36].

Bio marker Discovery
Early d iagnosis and treatment is a key factor required to reduce the mortality and morbid ity of all types of diseases, especially fo r tu mor disease. Unfortunately, currently available cancer and genetic screening tools (CT, X-ray, mammography and invasive needle or surgical evaluation for cancer and genetic disease) are not sensitive enough for early detection of the diseases thus most tumor diseases cannot be treated at an early stage. Moreover, it is imperative to develop non-invasive technique for diseases, such as tumors that can be distinguished between patients with and without cancer, as well as stages of cancer. At present, genomic and proteomic technologies have rapidly been involved in cancer research [37,38,39]. Geno mic technologies have allowed us to monitor thousands of gene expression profiles and evaluate functions of candidate genes to obtain a global view of cancer t issue. Proteomic techniques have also allowed us to understand proteins and their modificat ions. Despite its remarkable usefulness, microarray and proteo mics techniques all have technical limitations because of cell purity or limited cell nu mber after in itial purification fro m clin ical specimens so that, if we do not have a rat ional module fo r clinical genomic analysis, clin ical geno mic data cannot clearly define some specific bio markers and the accompanied therapeutic targeting.
In order to address the important question, here I introduce characteristics, advantages and disadvantages of the genomic analysis ex vivo, in vitro and in silico for application of clin ical specimen.
The most useful protocol uncovering biomarker is in vitro genomic analysis, such as single-cell man ipulation by laser capture microscopy (LCM) along with genomic analysis. If we have known cell mo rphology in a tu mor d isease and/or informat ion fro m cancer cell such as biomarker or telomerase activity, we can pick the cells with the morphological feature and enzy me activity, and then process genomic analysis fro m the harvested cells. After in vitro genomic analysis, we will uncover very specific biomarkers [40]. If the cells are located in liquid such as leukemia cells in bone marro w or peripheral b lood, living cell-sorting such as FACS/MACS with downstream genomic analysis allows us to discover specific bio markers fro m clin ical specimens.
Additionally, geno mic analysis ex vivo, or liv ing cell-sorting technique with the downstream cell culture to increase cell nu mber or understand differentiating stage and then genomic analysis can help us to discover some biomarkers at different stages of cancer [41]. As mentioned above, genomic analysis ex vivo has been employed into cancer stem cell.
Recently, genomic analysisin silico has been extensively utilized to discover tumor bio marker and bio markers related with genetics. As shown in our result [42], genomic analysis in silico can screen or mine some very specific bio markers in some diseases. Because of mixed different cells fro m clinical tissue remaining in the mined gene profile, current results fro m bio informat ic analysis still need further experiments, such as rtPCR or Western blot to confirm the mined gene profiles. Geno mic analysis in silico is a good substitute method if the specimens have not been processed for micro -dissection in vitro orcell-sorting in/ex vivo.

3.3.2.Therapeutic Targeting
Therapeutic targeting is a specific treat ment for clin ical diseases. In details, after genomic analysis along with pathway information and lin king drug database to provide informat ion, the corresponding drugs will directly target the explored therapeutic protein and nucleic acid for the targeted disease. A necessity of therapeutic targeting, clinical genomic analysis, nowadays, has been quickly applied into tumor disease, immune disease and genetic disease. Major developments are in two fields: A. therapeutic targeting for cancer stem cell; B. individualized therapy for metastatic tumor, immune disease or genetic disease. Now we all know, small populations of cells, called cancer stem cells (CSCs) located in tumor tissues play an important function in the development and progression of the disease [43,44,45]. It is also thought that CSCs drive the metastatic spread of cancer. CSCs are able to resist conventional therapies so that the disease is difficult to be completely erad icated. If therapeutic targeting identification can uncover some specific proteins or nucleic acid of CSCs, the selective targeting of CSCs will offer a new parad ig m in both cancer diagnostics and therapeutics. According to current report, mo re than 30 CSC R&D program are subject to several clin ical research to discover drugs or compound. Now, most CSC R&D program is being taken forward to large pharmaceutical co mpanies [46,47].
In clin ical fields, personalized medicine (or individualized therapy), one of special therapeutic targeting, is going to extend into different clinical diseases. Personalized medicine is a new medical model to be directly defined as physicians enable tailored approaches to prevent and care for indiv idual patient relying on genomics and proteomics. It is often defined as "the right treat ment for the right person at the right time." All examp les of successful personalized treat ments require a rat ional clinical geno mic analysis. Based on research and development of clin ical genomic analysis, we have successfully established a bioinformat ics module for personalized therapy, that is, fo llo wing genomic signature mined by genomic analysis in silico, quantitative pathway analysed by topology process, a specimen with no small cell lung cancer (NSCLC) was used to mine significant targeting, finally uncover therapeutic targeting with linking drug database for the special treat ment [48].

Conclusions of Clinical Genomic Analysis
Seven years ago, I describe single cell genomic techniques based on experiences in our laboratory and other laboratories. After further research and development of clin ical genomic techniques with maturity in amp lification for genomic DNA/ RNA and signal magnification for protein for the past several years, many new genomic techniques have been developed into clinical specimens, for instances, dynamically single-cell or small number of cell-sorting techniques with downstream ex vivo culture to increase the cell number or d ifferentiation into a special type of cells and then genomic analysis; moreover, geno mic analysis at tissue level with in silico analysis has also be quickly developed. These new clinical geno mic analyses have been extended fro m mo lecular pathology, mo lecular genetics/cytogenetics, forensic medicine into stem cells for regenerative medicine and tumor cells and/or cancer stem cells for bio marker discovery and therapeutic targeting and into T-cell for adoptive immunotherapy.