An Academic Report sets out the official stance of the Academy – the National Academy of Medicine of France (ANM) adopted the Report and Recommendations at its plenary session convened Tuesday February 23, 2016 by a vote.
Next generation sequencing (NGS) technologies, which allow high speed automated DNA sequencing, DNA sequence analysis and comparisons using big data algorithms are being used more and more in medical diagnosis, to provide prognosis information and for choosing a treatment that best fits the patient.
The French “Académie nationale de médecine” (ANM) and “Académie des technologies” (NATF) jointly stress the scientific and medical importance of these technologies and call for bringing together strengths of the national computer science industry and the medical community in a public-private consortium in order to build a Demonstrator and Centre network filling up the growing gap between France and the most advanced countries in this field with great potential. They underline also common pitfalls which need to be addressed in technic, legal, economic, education and ethics issues, in order to make this complex project a success.
A great many examples demonstrate the relevance of this approach. This category of medical analysis is already used in hospital services round France to screen, identify and monitor “rare diseases” and also as part of cancer treatment protocols for which purpose the National Institute for Cancer Research (INCa) set up a nationwide organization comprising 28 platforms for molecular genetic tumour studies. In this way, the screening process to identify genetic mutations responsible for hereditary diseases is made easier and also serves to throw light on the extreme heterogeneity of tumours that can affect one type of organ, this leading to specifically targeted therapies, as for example in the case of stage 4 colorectal cancer, where mutations of the K-RAS gene render treatment by the monoclonal anti EGFR antibody inefficient, whereas for a metastatic melanoma, in reverse, the V600E mutation of the BRAF gene can indicate a case for using verumafinib [a B-Raf enzyme inhibitor] treatment.
These techniques are also widely used in research activities, whether it be to gain a better understanding of the relationships between phenotypes and genotypes in the two sorts of pathology just mentioned above, but also in discovering their variants (“single nucleotide polymorphism” or SNP) present in common non-transmissible diseases. The question then is not “Should we develop a genomic approach in medicine?” but rather “How do you proceed from a targeted analysis of a few genes to the whole exome, or better still, the complete genome?”, given that only the sequencing of the complete genome allows you to have a full picture of the exome. Moreover, analysis of the complete genome offers considerably more and paramount information about gene regulation than that of the exome alone and is of better quality, even if at the present time we do not yet know how to analyse and understand all of them. Furthermore, in parallel to the development of personalized medical care, if we had sufficient capacity to carry out large-scale sequencing genome screening, it could be used as a substitute for amniocenteses that may cause abortions or present risks to the foetus, i.e., by analysing foetal DNA that is present in the mother’s bloodstream.
France is definitely lagging behind in the area of NGS and does not have national high through-put sequencing infrastructures capable of carrying out thousands of analyses each year, as is the case for certain other countries e.g., the USA, China or the United Kingdom which have already created (or are in the process of committing themselves to) such infrastructures.
Inasmuch as the ANM and the NATF national academies in France were fully aware of the acceleration in implementation of these technologies, they decided to appoint and commission a joint working party in 2014 to assess these developments with the remit to make recommendations related to designing and building such structures, including the some of the necessary organizational, regulatory, technological, economic, educational and ethical factors, as described hereinafter:
The proposal made by the AVIESAN group, based on data provided by INCa suggests the creation of between 10 to 15 centres each having an analytical capacity comparable to that of an Xten Illumina sequencer, viz., 19 000 whole human genomes analysed every year. In this organisation scheme proposed by this Group, each sequencing centre possesses the capacity to analyse its own data, for the purpose of establishing a n initial report to the practitioner that details not only the relevant sequences for the investigation, but also offers an interpretation by way of a diagnosis, prognostics and possible treatment all of which constitutes an aid to a medical decision process under the sole responsibility of the doctor in charge of the patient or of a specialist.
The patient-practitioner relationship is vital inasmuch as – inter alia – it enables the patient (and possibly his/her ascendants and descendants who do not have the disorder in question (heterozygotes)) to understand and to accept the rationale of a genome analysis, and also to make sure that the ill person and his close family members correctly understand the results of the analysis and their possible implications. In the scheme proposed all the sequencing centres are inter-connected and inter-operational. Moreover, they would be connected to a centralised data processing centre that collects all the genomic data to optimise the analyses but also collect all the other clinical data related to the patient, even including data from various connected objects that are rapidly developing today, the complete picture constituting the patient’s personal medical file.
The core objectives of the central data processing unit are both immediate and long-term: they are immediate in terms of the interpretation given to the genomic data plus other clinical data, and long-term to the extent that the centralised data will be preserved and thereby build up a rich store of centralised data that will prove very valuable for research work and, as the advancement pf knowledge progresses in this field, might be re-interpreted.
Both of our Academies agree with this organisational scheme. They nonetheless point out that there are numerous, dissimilar problems arising from a change in scale of genomic analysis implied by personalised medical care, as we mentioned earlier.
We shall now examine, non-exhaustively, these problems which led the academic Working Party to recommend that in the first instance a demonstrator be designed and built, in order to test the various solutions possible before full national implementation of the scheme as proposed by AVIESAN.
1 – DNA sequencing
Today, the world scene of DNA sequencing units is largely dominated by an American company, Illumina Inc., who consequently enforce their sequencing technology, their data digitising algorithms, the unit prices their equipment and also those for the reactives. This as yet undisputed monopoly enables the company to establish (and enforce), as the standard in the field, its sequencing technologies and protocol as well as its data storage procedures for which Illumina Inc. will be the official owner of the rights and, beyond these considerations, the methods to be used for diagnosis purposes.
There is very little chance in the near future of seeing this situation cease because the alternative solutions, such as that envisaged by Complete Genomics, a subsidiary of the Beijing Genomics Institute (BGI, Chine) seemingly are abandoned, for the moment. And the situation will probably continue unchanged for several years. A technological watch-tower operation is needed to learn as early as possible about possible new projects that could put an end to Illumina’s monopoly. In the meantime, we should be seeking to establish the bases for a long-term contract with Illumina Inc. Such an agreement however, should not mask or obviate the conditions needed to create a sufficiently robust and viable downstream market, so as to be able to launch a real genomic data interpretation software industry. The example that come to mind is “Genomics England”, a project which began in the UK in 2012 (described below in the Appendix); it merits all our attention to see how the British chose to solve this essential question.
Notwithstanding, it can be noted that the development of techniques used before and after the actual sequencing phase remain open to competition and could readily be developed by innovative start-ups, upstream in the area of micro-fluid mechanics to prepare samples for sequencing and associate reactive agents – fields in which the skills and knowledge already exist in France (e.g. at the CEA-LETI Lab., at ESPCI, etc., and downstream in the fields of genomic data analysis and storage.
Lastly, we can only hope that with the decade or so, sequencing itself will no longer be limited to Illumina‘s Xten type machines and that a French sequencing sector, or European for that matter, could be developed .
2 - Digital processing of digitised data derived from DNA sequencing operations
Today, while it is commonplace to carry out DNA sequence comparisons, the processing of very large quantities of genetic sourced data via rapid analysis protocols with results made accessible to the practitioners is an evolving scene. For a very large computer centre such as that at the CEA-TGCC installation, France disposes of tools such as the CCRT (450 teraflops) research unit which is open to industrialists or the Curie (2 petaflops) unit used for European level research projects, financially supported by the Genci programme (acronym in French for national large scale intensive computation systems), operated by the CEA.
The Teratec, a European cluster of skills in high-level modelling which brings together various private companies working in the digital analysis field (and which houses the TGCC) has actors who are fully competent and are capable of handling programmes such as ATOS or INTEL both in terms of software and hardware skills. A few start-ups, specialist in this kind of analysis, such as Biofacet or Intragen seems to be interested in participating in the technology-intensive challenge, including the software lead company Dassault Systems. The CCRT already holds digitized sequences for the France Genomics Group and could serve as a data storage node for data forthcoming from the proposed Demonstrator. Moreover, the development of algorithms specifically targeting medical and biological data (in a wide connotation) is also necessary.
These advantageous factors, industrial as they are in essence, must necessarily be reinforced in the framework of personalized medical care. It is obvious that future challenges will not lie in the sequencing but more in data storage and analysis (involving confidentiality and sensitive information about patients) as well as being important for research, for diagnosis and choice of treatments.
In terms of regulatory texts
It is also important to separate research-oriented activities from diagnosis, inasmuch as the regulations applicable will differ in these two approaches.
As far as very large scale sequencing is concerned (for the exome or whole genome) in the framework of research activities, the regulations applicable are somewhat complex and those that exist today are not exactly adapted to the cases. This problem is made even more complex when the genetic characteristics of persons are analysed and recorded.
Today, biomedical research can only be conducted after an “in favour” decision of the special committee for the protection of persons (CPP in French) and the computerized files must be declared to the French national committee for computer sciences and liberties (CNIL). All participants must sign a document to indicate their informed consent. The precise nature of the files is not specified in the law, and this removes problems concerning nucleic acid DNA sequences (if the CPP above declares them to be justified)? This regulatory context makes it difficult to carry out systematics screen studies of human genomes which by definition do not normally relate to physio-pathological hypotheses; the objective of the studies is to show an association between a given phenotype (generally, this refers to clinical symptoms that cannot be attributed to a particular pathology) and genome sequences.
Moreover, there are medical data collected over numerous years screening of millions of individuals stored on the computers of the French ministry in charge of Health.
They represent a major source of data for any research in terms of personalized medical care. This data storage is perfectly legal per se. In contradistinction, the authorization to compare medical data with genome data will prove much more difficult to obtain. To conclude, even it at the moment there is nothing to legally prevent studies with a defined pathology in mind, it will probably be necessary to introduce changes in legislation for major personalized medical care programmes. A serious amount of work and collaboration will be needed with the CNIL committee who are especially vigilant when it comes to comparing and cross-analysing files. Storing data for nucleic acid sequences does not raise any technical problems or issues. Several private companies already offer this service and have been duly authorized to do so. The factors that enable authorization rely mainly on a demonstrated safe conservation of the data and its confidentiality. Implementing very large-scale storage infrastructures in France should not consequently encounter any difficulties.
As far as diagnostic activities are concerned, the regulations currently enforced are clear, especially since the law known as “Hospitals, health and territory (HPST)” was adopted in Parliament, which initially covered the “biological” aspects in the executive order paper N°2010-49, January 13, 2010 on medical biology – the so-called Ballereau order. Any diagnostic test is considered as a medical act which can only be performed in a bio-medical centre (laboratory) by a practitioner, a doctor or a chemist holding a specialist diploma (DES) in medical biology. These laboratories must be certified by the French centre for accreditation (COFRAC). Wherever genetic tests are carried out, various other regulations apply, defined in the successive laws on bio-ethics, the latest version of which dates back to 2011 and which very precisely sets out the procedures for prescription of the tests, their implementation and the way in which the results are made known to the patient. Any structure that aims at conducting large scale tests must therefore be a medical biology laboratory, whether it be a public or a private structure, in compliance with current regulations.
This restriction to activities will only become problematic later, since a medical test can only be proposed to a patient if and only if its medical value notably its predictive results, including both negative and positive aspects, is high and totally demonstrable. A considerable amount of work remains to be done here in order to have these perquisites attainable, thereby enabling a test to be proposed for personalized medical care for commonplace diseases. Given the prevailing context today– but maybe probably for some time to come, this will not extend to Whole exome sequencing (WES) or to whole genome sequencing (WGS) which are in the assignment by the Prime Minister (cf. footnote 1).
The decision taken last summer by the ministry in charge of Health entrusting the cost of genome sequences less than 500 kb to the Social Security system illustrates the limitation well, excluding de facto (and intentionally) both WES and WGS testing which remain considered as research and not diagnostic activities. If new tests, compatible with the specifications of medical biology, become operational, the question of their implementation should be resolved fairly easily, since there are more than 100 laboratories in France who have the required skills and knowledge bases. The problem will lie with the equipment available and the question as to whether it would prove better to mutualise the equipment among a small number of vert large-scale structures. The law as it stands today (regulating all medical biological practices) would become a serious handicap to progress and it would require adaptation if the option above were to be take, seriously.
In economic terms
The changeover from small-scale centres such as those currently being developed by INCa, limited to analysing panels of genes, to vert large-scale centre appropriately equipped to the point that they would generate important economies of scale for the cost per base sequenced and analysed, if there are preferential agreements associated with the use of Xten type machines. From this point of view, the creation of large-scale centre is economically logical and necessary if as is announced in in the National Cancer Plan, an analysis of 50 000 whole genomes by year 2019. However, the increase in screening demands induced by the existence of these large centres will lead to a high overhead cost. In reverse, having the capacity to make better diagnosis and prognosis associated with a reasoned choice he it comes to a better adapted therapy will not only lead to the patient’s having better morale and medical benefits, will also provide for substantial cost savings. In attempting to clarify these areas of uncertainty, we simply cannot avoid analysing economic return on the investments (ROI) needed to implement such a network of large centres as proposed.
The Demonstrator we have in mind could be included in an application lodged with the General Commissariat for Future Investments (CGI). Should this policy option be adopted, we would have also to ensure the financial participation of industrialists and/or financial partners who view the platform as an opportunity to test various technological solutions and/or business models. An assessment of the economic value that would stem from this sort of large-scale medical screening and analysis could be launched in the health sectors, to measure the potential input from, e.g., pharmaceutical companies, health sector mutual insurance companies, computer science and applications companies, etc.
Logically speaking, the sheer mass of data interconnected in this way should enhance the medicinal drug sector through development of new therapeutic molecules deemed suitable for appropriate action on the human genome (pharmaco-genetics) and likewise for the development of genetic tests.
To ensure better levels of profitability for the future computation Centre, we could imagine using the computers available to carry out analyses of DNA samples from other sectors, for example, from plant or animal sources, where we foresee some very important applications for breeding and selection purposes. France has some companies who are already world class leaders in plant selection and animal breeding.
In terms of education
Development of personalized medical care as outlined here can only be conceived with an accompanying development of training that includes medicine, biology, computer sciences in all sorts of format, with a closer tie between the Universities and the French Engineering Schools (grandes écoles). All the experts consulted underline the cruel shortage of skilled personnel in the area of bio-computing and data processing, which shortfall must be settled in a competitive manner before the network envisaged here is deployed.
In terms of ethics
The ethical questions that arise through forecast uses of NGS extend the range of existing questions that stem from today’s classic genetic analyses and protocols.
1- Individual genome analyses and uses made of data. Genome analysis allows for a perfect identification of a person and gives valuable indications about existing disorders and potential disease risks. This information is personal and therefore cannot be used without the prior consent of the person tested, except when there is a legally authorized injunction taken in the course of a police enquiry. No practitioner, laboratory or institution whatever can sequence, analyse and preserve genome samples unless the person involved has given consent in writing and furthermore has been properly informed about the use(s) that will be made of his/her data. In the case of children, analysis is possible with the parents’ consent if and only if there is an immediate interest for the child or a member of the family to make decisions that can be preventive or therapeutic. One outstanding problem is that of how to report the analyses.
Given that a whole genome analysis allows you to discover genetic mutations or polymorphisms in thousands of genes provides information, that can prove useful when it comes to choosing among possible preventive treatments, and identifies the precautions to be taken in the case of administration of medicinal drugs and also, in the case of recessive mutations, to aid the decision of a couple to employ (or not) a prenatal diagnosis. It is not uncommon for the practitioners to discover various anomalies that were not predicted over and above the data being investigated. Most doctors agree that the patient must be informed correctly in advance as to the possibility of such discoveries, of the deficit often to interpret their significance in particular when the penetration level of the mutation is variable, and to offer medical advice to the patients.
They likewise agree on the fact that it is important to clearly inform the patients of these unexpected results when treatment or preventive measures are proposed. French legislation today limits the possibilities of carrying out genetic analysis to medical practice, research and to assist in legal enquiries. In reality, anyone can obtain a genome analysis via a foreign laboratory. The results in this case are forwarded to the patient, together with medically oriented comments, and mentioning genetic polymorphisms that can indicate above average risks for various chronic illnesses. In most cases, this serves only to increase the patient’s anxiety. It is then the practitioner’s role to reassure the patient as to the probabilistic nature of the data and their low predictive impact.
2- Preconception and prenatal testing. Preconception tests are not often taken in France, since the chances of a man and a woman with the same recessive genetic mutation who wish to procreate is very small. The question will, however, arise in the case of a global genetic screening test programme on a population. In such circumstances, rather than give the individuals in the population all the information obtained through screening, including the observation of recessive or dominant mutations with clinical symptoms that will become observable in the long term, it would be better to limit the information to that with immediate medical consequences that call for treatment or preventive measures. Classic prenatal diagnosis or pre-implant genome analysis of very early stage embryos are permissible when there is a risk of a severe genetically inherited disorder in order to make choices among ‘clean’ embryos for the uterine implant operation. In such cases, the genome investigation must be restricted to seeking the mutation that triggers the illness in question, to avoid any inappropriate practice.
3- The project to build a national genome data base. There already are data bases in numerous countries, including France, to store the DNA of those with criminal records, this enabling possible attribution of criminal offences to individuals “on record” in the French National Automated Genetic Print File (FNAEG). A new step is currently under way to build u national data bases for large numbers of individuals. There are multiple advantage in having such bases. They can be a prime epidemiological research tool to discover and quantify in certain populations the existence of various mutations and polymorphisms and, in the long term, to relate these phenomena with certain ethnic origins. The bases could also allow research practitioners to associate as yet unknown mutation to given phenotypes. Lastly, it allows the authorities to extend the national file system with data that supersedes today’s neonatal screening tests, that concern only a small number of medical, genetic disorders. Personal sequencing data can be stored for a lifetime, as a sort of “genetic ID card”, that might even be included on the French ‘carte vitale’ (today’s social security medical care card) or its equivalent. Over and above the fact that the associate costs would not be negligible, there are also those ethical issues (mentioned earlier) that would stem from possible communication to parents of an as-yet not apparent disorder for their child, but which ailment might possibly be severe in the long run.
Conclusion and Recommendations
The joint Academic Committee (National Academy of Medicine (ANM) and the National Academy of Technologies of France (NATF)) came to the conclusion – if France is to preserve its excellent position as personalized medical care policies progress - that it is now necessary to develop as rapidly as possible at a national level those genome analytic capacities needed, by using the latest generation of sequencing technologies available.
Indeed, the hearings conducted by the joint Working Party representing both academies underlined the fact that France is falling behind compared with other countries such as the USA, China or the UK in the mastery of the technologies both from a purely sequencing point of view and in the fields of data processing and storage. Another working party, appointed by Aviesan (cf. footnote N°1) also very clearly demonstrates this weakening position in its findings.
Inasmuch as we are fully conscious that the long term development of the analytic capacities needed raises numerous questions (given the sheer mass of data to be processed), in terms of organization, regulatory texts, technologies, economics, finance, education, training and ethic, we strongly recommend that a first step should consist of setting up a demonstrator unit to test on a scale one those solutions available and seen as possible.
By way of contrast, we can, however, imagine that the change in scale in genome analyses, as we have outlined in this Report, will offer a major opportunity for the development not only of medical practice, but also in the industrial sectors focused on medicinal drugs and diagnostic test protocols. Thus, the Academies consider that whereas the sequencing needed today will for the near future necessarily require the services of the American units provided by Illumina Inc., French industrial and research skills and know-how will be able to compensate this setback, notably through very high level computer science applications on a national scale.
Following suit to an analysis of the situation in France, both academies ANM and NATF issue the following recommendations among which the fourth (#4) and fifth recommendations (#5) presuppose that the current legislation be modified.
- To set up a demonstrator unit sufficiently large that it can carry out some 40 000 whole genome analyses per year, can assure their digital format storage and interpretation in associating the largest number of “user” partners possible who will thereby contribute to its added value.
- To encourage the creation of a French (or European) company to assemble sequencing units with an equivalent (or superior) capacity than that proposed by Illumina Inc., in order to bring the monopoly of this US company to an end.
- To develop, upstream of the sequencing operations, micro-fluidic approaches to prepare the samples for analysis and downstream, the data interpretation software needed by setting up the contacts needed between industrialists and geneticists.
- To modify current rules and regulations, to enforce the fact that all research projects that implying genome sequencing be dependent on a physio-pathological hypothesis to the extent that the objective sought by systematic screening sequencing is to establish this very relationship.
- Under the control of the French national committee for computer sciences and liberties (CNIL), to facilitate for research purposes only, the cross-referencing of genetic data files with clinical files related to the same, given patients, whilst respecting their anonymity.
In a second phase, to foresee and plan the creation of between 10 to 15 Centres located round the country, inter-connected and inter-operational, plus one centre specifically dedicated to collecting not only all the genetic data from the sequencing operations, but the entire range of data generated concomitantly.
THE GENOMICS ENGLAND DEMONSTRATOR PROJECT
The “100.000 genome” programme of the Department of Health, launched by the British Prime Minister end 2012 is a good example pf public/private partnership in the specific framework of State run health services. This programme has already become operational.
The programme missions and objectives were clearly set out at the start of the project:
• To create a genomic medicine service within the NHS (National Health Service) bringing benefit to patients
• To create an ethical and transparent programme based on consent of participants
• To enable new scientific discovery and medical insights
• To kick-start the development of a UK genomics industry
The project called for the setting up of a private company “Genomics England”, financed through public funding sources with various stakeholders in the Health Department: the National Institute for Health Research, NHS England, Public Health England and Health Education England.
Initially focusing on England, the company could integrate other public parties from the United Kingdom. North Ireland recently joined the ranks of the Project.
Fully aware of the difficulties to implement such a project in an ever-moving technological framework, Genomics England have decided to progress by stages. The first step consists of building a ‘demonstrator” focused exclusively on two medical problem areas:
• rare diseases.
The demonstrator envisages aims at analysing 70 000 DNA samples (100 000 if we take into account multiple biopsies from a single patient in the case of cancer metastases).
The samples will be taken on an informed consent voluntary basis and the company has committed itself to guarantee that the results – even when made anonymous – will not be forwarded to interested industrial partners without prior and explicit consent.
Genomics England signed a contract in 2014 with the American company Illumina Inc. to carry out the sequencing operations. The agreement has a public financial support amounting to 78 m£; Illumina Inc. will be investing 162 m£ in England over a 4 year period for this project.
On an operational level, Genomic England will benefit also from an investment of 27 m£ for a building to be erected on the Welcome Trust Campus, near Cambridge. The Sanger Institute, also financially supported by the Trust, is a party to the sequencing operations.
(ii) Genomic data processing
The data collecting and processing aspects will be partly funded by the Medical Research Council for an amount of 24 m£.
Very recently - (November 2015), Genomics England was chosen and contracted with WuXi Next Code, an American subsidiary of the Chinese Group WXi App Tec to ensure clinical interpretation of the sequencing results. This private company is itself a spin-out of deCODE Genetics.
The company has recently announce that it has chosen Cognizart to help define and implement an operation computer and data processing environment. Cognizart is an American EDP specialist company working, inter alia in the health and pharmaceutical fields.
NHS England has programmed a 20 m£ budget for the duration of the project to finance a call for project to identify the future genomic medicine centres for the NHS. To date, some 28 public research teams have been selected. They are integrated into what is known as the “GENE Consortium”, alongside 10 private pharmaceutical and biotechnology-intensive companies who will also make financial contributions to the overall project.
The highly operational organization described here illustrates the choice that rather than progressing solely on the basis of public research, the option was made from the start to bring in some of the best private sector companies as partners , including some foreign partners, to fulfil the objectives within a limited time horizon.
The level of public investment, as per the figures communicated, totals 129 m£, with Illumina Inc. investing some 162 m£ (over a 4 year period, covering the access rights to the latest generation machines, for the reactive agents and for the specialist personnel need to run the operations).
Ethics monitoring of the programme is ensured by nomination of a Committee the assigned role of which will be to supervise the reporting of the genomic analyses and, in close liaison with the NHS, to draft the associated regulatory texts and system.
Signed and certified
The Permanent Secretary
Prof. Daniel COUTURIER