Method for formulating biomarker reference interval based on transcriptome sequencing gene expression quantity data

Abstract

The invention discloses a method for formulating a biomarker reference interval based on transcriptome sequencing gene expression quantity data. The method comprises the following steps: constructing transcriptome sequencing gene expression quantity data sets of a healthy group and a patient group; screening out low-expression genes in a healthy group and a patient group; detecting and removing outlier samples in the data sets of the health group and the patient group based on an IQR method, and filling outlier points; extracting counts numerical values of corresponding genes according to common gene names in the transcriptome sequencing gene expression quantity data sets of the healthy group and the patient group; the gene length and the sequencing depth are standardized, and the count value of the extracted gene is converted into a TPM value; determining a reference interval based on an SMR algorithm; determining the trend of each gene; and determining a target biomarker and a corresponding reference interval. According to the method, the skewed distribution characteristics of transcriptome sequencing gene expression quantity data can be well matched, and the reference interval of the biomarker gene expression quantity is divided.

Description

Technical Field

[0001] The present invention relates to the technical field of biological data reference interval determination, and in particular to a method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data. Background Art

[0002] In the biomedical field, it is often necessary to test multiple indicators to evaluate the physiological condition of biological individuals. At this time, a reference interval is often needed to help doctors or researchers determine whether the sample test value is abnormal.

[0003] There are two commonly used methods for establishing reference intervals:

[0004] 1. Indicates the 95% data distribution range;

[0005] 2. Determine the reference limits of 2.5% and 97.5% by percentile method to determine the 95% reference interval, or use the lower limit of P5 or the upper limit of P95 unilaterally to determine the reference interval;

[0006] Method 1 requires that the data satisfy the normal distribution; the reference interval defined by method 2 may not be the interval with the most concentrated samples. For the skewed distribution of transcriptome sequencing gene expression data, the above two methods are not well suited for the division of reference intervals. Summary of the invention

[0007] In order to overcome the defects and shortcomings of the prior art, the present invention provides a method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data. The present invention can divide the reference interval of biomarker gene expression based on the skewed distribution characteristics of transcriptome sequencing gene expression data. This interval is the interval where the gene expression levels of 95% healthy samples are most concentrated.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] The present invention provides a method for formulating a biomarker reference interval based on transcriptome sequencing gene expression data, comprising the following steps:

[0010] Obtain transcriptome sequencing data and construct transcriptome sequencing gene expression datasets for healthy and patient groups;

[0011] Setting screening conditions for low-expression genes, and screening the transcriptome sequencing gene expression data sets of the healthy group and the patient group based on the set screening conditions;

[0012] Based on the IQR method, outlier samples in the transcriptome sequencing gene expression dataset of the screened healthy group and patient group were detected and removed, and outlier points were filled;

[0013] Extract the count values ​​of the corresponding genes based on the common gene names in the transcriptome sequencing gene expression datasets of the healthy group and the patient group;

[0014] Gene length and sequencing depth were standardized, and counts of extracted genes were converted into TPM values;

[0015] Based on the SMR algorithm, the interval with the most concentrated distribution of TPM values ​​of 95% of the samples in the healthy group transcriptome sequencing gene expression data set was obtained as the reference interval;

[0016] The trend of each gene was determined based on the TPM value of the transcriptome sequencing gene expression dataset of the patient group;

[0017] The target biomarkers and corresponding reference intervals were determined based on the gene trend results.

[0018] As a preferred technical solution, the transcriptome sequencing gene expression datasets of the healthy group and the patient group are screened based on the set screening conditions, specifically including:

[0019] For each gene, the proportion of counts values ​​less than the set value is calculated, and the expression values ​​corresponding to genes with a proportion greater than the set proportion are screened out.

[0020] As a preferred technical solution, outlier samples in the transcriptome sequencing gene expression dataset of the screened healthy group and patient group are detected and removed based on the IQR method, specifically including:

[0021] For each gene remaining in the healthy group and the patient group after screening, the data outside the interval of [Q1-1.5IQR, Q3+1.5IQR] were screened out, where Q1 was the 25% quantile, Q3 was the 75% quantile, and IQR=Q3-Q1.

[0022] As a preferred technical solution, the filling of outliers specifically includes:

[0023] The average expression of transcriptome sequencing genes after removing outlier samples was calculated and the average value was filled in the outlier points.

[0024] As a preferred technical solution, based on the SMR algorithm, the interval with the most concentrated distribution of TPM values ​​of 95% of the sample size of the healthy group transcriptome sequencing gene expression data set is obtained as the reference interval. The calculation formula for the lower and upper limits of the reference interval in the SMR algorithm is expressed as follows:

[0025] (L i , U i )={(min(T ik ), max(T ik))|min(ΔT ik )}

[0026] ΔT ik =max(T ik )-min(T ik )

[0027] Among them, T ik represents the kth subset of the set of TPM values ​​of all samples of the i-th gene in the healthy group, X ij represents the TPM value of the jth sample of the ith gene in the healthy group, j∈[1,n], n represents the total number of samples in the healthy group, and the subset T ik The number of elements k∈[1,a], a is the number of combinations card() indicates the number of elements in the collection. It means that among n samples, 95% of n samples are rounded down and ΔT ik Represents a subset T ik The difference between the maximum and minimum values, L i Indicates the lower limit of the reference interval, U i Indicates the upper limit of the reference interval;

[0028] The healthy group was obtained by SMR algorithm The interval where the TPM values ​​of samples are most concentrated [L i , U i ] as the reference interval.

[0029] As a preferred technical solution, the trend of each gene is determined based on the TPM value of the transcriptome sequencing gene expression data set of the patient group, specifically including:

[0030] Score each sample for the patient group:

[0031]

[0032] Among them, Y ih represents the score of the hth sample of the i-th gene in the patient group, S ih represents the TPM value of the hth sample of the ith gene in the patient group;

[0033] Determine the trend for each gene:

[0034]

[0035] Among them, A i is the trend result of gene i, and N represents the total number of samples in the patient group.

[0036] The present invention also provides a system for formulating biomarker reference intervals based on transcriptome sequencing gene expression data, including: a data set construction module, a gene screening module, an outlier data removal module, an outlier filling module, a gene counts value extraction module, a TPM value conversion module, a reference interval construction module, a gene trend calculation module, and a result output module;

[0037] The data set construction module is used to obtain transcriptome sequencing data and construct transcriptome sequencing gene expression data sets for the healthy group and the patient group;

[0038] The gene screening module is used to set screening conditions for low-expression genes, and screen the transcriptome sequencing gene expression data sets of the healthy group and the patient group based on the set screening conditions;

[0039] The outlier data removal module is used to detect and remove outlier samples in the transcriptome sequencing gene expression data set of the screened healthy group and patient group based on the IQR method;

[0040] The outlier filling module is used to fill the outliers;

[0041] The gene counts value extraction module is used to extract the counts values ​​of corresponding genes according to the common gene names in the transcriptome sequencing gene expression data sets of the healthy group and the patient group;

[0042] The TPM value conversion module is used to standardize the gene length and sequencing depth, and convert the counts value of the extracted gene into a TPM value;

[0043] The reference interval construction module is used to obtain the interval with the most concentrated distribution of TPM values ​​of 95% of the sample size of the healthy group transcriptome sequencing gene expression data set based on the SMR algorithm as the reference interval;

[0044] The gene trend calculation module is used to determine the trend of each gene based on the TPM value of the transcriptome sequencing gene expression data set of the patient group;

[0045] The result output module is used to determine the target biomarker and the corresponding reference interval based on the gene trend result.

[0046] The present invention also provides a computer-readable storage medium storing a program, which, when executed by a processor, implements the method for formulating a biomarker reference interval based on transcriptome sequencing gene expression data as described above.

[0047] The present invention also provides a computer device, comprising a processor and a memory for storing a program executable by the processor, wherein when the processor executes the program stored in the memory, the method for formulating a biomarker reference interval based on transcriptome sequencing gene expression data as described above is implemented.

[0048] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0049] The present invention adopts the innovative SMR (search for the minimum range) algorithm, which can divide the reference interval of biomarker gene expression for the skewed distribution data of transcriptome sequencing data. This interval is the interval where the gene expression of 95% healthy samples is most concentrated. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The present invention is a flow chart of a method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] Example 1

[0053] like Figure 1 As shown, this embodiment provides a method for formulating a biomarker reference interval based on transcriptome sequencing gene expression data. After obtaining the healthy group / patient group transcriptome sequencing gene expression data set, the two groups of data are preprocessed by screening low expression, removing and filling outliers; the counts values ​​of the corresponding genes in the two groups are extracted according to the common gene names of the healthy group and the patient group; the gene length and sequencing depth are standardized to obtain the TPM value; the TPM value reference interval is delineated by the SMR algorithm; the gene trend is determined by the proportion of the number of patient group samples in the reference interval; the target biomarker is screened according to the trend and the final reference interval is obtained.

[0054] The transcriptome counts data in this example is affected by gene length and sequencing depth and is not suitable for comparison between samples. Therefore, after standardizing the gene length and sequencing depth, the TPM value is used to define the reference interval. The specific steps include:

[0055] S1: Acquire transcriptome sequencing data, which is to obtain transcriptome sequencing gene expression datasets of healthy people and patients from public databases or clinical samples;

[0056] S2: Screen out low-expression genes. The specific steps are as follows:

[0057] Transcriptome sequencing will obtain a large number of short fragments, called reads. Counts is the number of reads aligned to each gene position on the reference genome. It reflects the original expression value of the gene in the sequenced sample. For each gene, the proportion of counts values ​​less than the set value is calculated. The set value of this embodiment is preferably 10, and the expression values ​​corresponding to those genes whose proportion is greater than the set proportion value are screened out. The set proportion value of this embodiment is preferably 50%. The above operations are performed on the healthy group and patient group data respectively. This type of low-expression gene is not easy to detect and is not suitable as a biomarker. And screening out low-expression genes can make subsequent calculations more efficient.

[0058] S3: Detect and remove outlier samples in the data set based on the IQR method. The specific steps are as follows:

[0059] After filtering out low-expression genes, for each remaining gene in the healthy group / patient group, samples outside the interval [Q1-1.5IQR, Q3+1.5IQR] are filtered out. Quartile data: Q1 (25% quantile), Q3 (75% quantile), IQR (Q3-Q1), outliers may cause deviations in subsequent calculations, so they need to be detected and removed.

[0060] S4: Fill in outliers. The specific steps are:

[0061] The average gene expression of the remaining samples after the outliers are removed from the healthy group / patient group is calculated, and the obtained average value is filled into the outliers. After the outliers are removed in step S3, missing values ​​are generated and the number of samples is reduced. In order to keep the number of samples of each gene consistent, the missing values ​​are filled with the average value.

[0062] S5: Extract the counts of the corresponding genes in the healthy group and the patient group based on the gene names shared by the two groups;

[0063] S6: Standardize the gene length and sequencing depth, and convert the gene counts extracted in step S5 into TPM values. Transcriptome counts data are affected by gene length and sequencing depth and are not suitable for comparison between samples. Therefore, the TPM value is obtained after standardizing the gene length and sequencing depth. The subsequent reference interval delineation and target biomarker screening are calculated using the TPM value.

[0064] S7: The reference interval is determined using the original SMR (search for the minimum range) algorithm. The lower limit of the reference interval is L i With upper limit U i The calculation formula is as follows:

[0065] (L i , U i )={(min(T ik ), max(T ik ))|min(ΔT ik )}

[0066] ΔT ik =max(T ik )-min(T ik )

[0067] Among them, T ik represents the kth subset of the set of TPM values ​​of all samples of the i-th gene in the healthy group, X ij represents the TPM value of the jth sample of the ith gene in the healthy group, i∈[1,m], j∈[1,n], m represents the number of genes in the sample, and the number of genes in each sample is equal, the subset T ik The number of elements k∈[1,a], a is the number of combinations card() indicates the number of elements in the collection. It means that among n samples, 95% of n samples are rounded down and ΔT ik Represents a subset T ik The difference between the maximum and minimum values;

[0068] The SMR algorithm can be used to obtain the health group The interval where the TPM values ​​of samples are most concentrated [L i , U i ] as the reference interval.

[0069] S8: Determine the trend of each gene. The specific formula is:

[0070]

[0071] Among them, Y ih is the score of the hth sample of the i-th gene in the patient group, S ih represents the TPM value of the hth sample of the i-th gene in the patient group, i∈[1,m],h∈[1,N];

[0072]

[0073] Among them, A i is the trend result of gene i, and N represents the total number of samples in the patient group;

[0074] S9: Determine the target biomarker and the corresponding reference interval, A iWhen the value is 1, it means that more than 95% of the TPM values ​​of N samples in the patient group are outside the reference interval, and from step S7, it can be seen that the healthy group has The TPM values ​​of the samples are within the reference range, and the TPM value distributions of the two groups are clearly distinguished, indicating that this gene can be used as a biomarker, and the reference range of the TPM value of this biomarker determined in step S7 is [L i , U i ].

[0075] The present invention adopts the SMR algorithm to divide the reference interval of biomarker gene expression for the skewed distribution data of transcriptome sequencing data. This interval is the interval where the gene expression of 95% healthy samples is most concentrated, which can well match the skewed distribution characteristics of gene expression in transcriptome sequencing data.

[0076] Example 2

[0077] This embodiment provides a system for formulating biomarker reference intervals based on transcriptome sequencing gene expression data, including: a data set construction module, a gene screening module, an outlier data removal module, an outlier filling module, a gene count value extraction module, a TPM value conversion module, a reference interval construction module, a gene trend calculation module, and a result output module;

[0078] In this embodiment, the data set construction module is used to obtain transcriptome sequencing data and construct transcriptome sequencing gene expression data sets for the healthy group and the patient group;

[0079] In this embodiment, the gene screening module is used to set screening conditions for low-expression genes, and screen the transcriptome sequencing gene expression data sets of the healthy group and the patient group based on the set screening conditions;

[0080] In this embodiment, the outlier data removal module is used to detect and remove outlier samples in the transcriptome sequencing gene expression data set of the screened healthy group and patient group based on the IQR method;

[0081] In this embodiment, the outlier filling module is used to fill the outliers;

[0082] In this embodiment, the gene counts value extraction module is used to extract the counts values ​​of the corresponding genes according to the common gene names in the transcriptome sequencing gene expression data sets of the healthy group and the patient group;

[0083] In this embodiment, the TPM value conversion module is used to standardize the gene length and sequencing depth, and convert the counts value of the extracted gene into a TPM value;

[0084] In this embodiment, the reference interval construction module is used to obtain the interval with the most concentrated distribution of TPM values ​​of 95% of the sample size of the healthy group transcriptome sequencing gene expression data set based on the SMR algorithm as the reference interval;

[0085] In this embodiment, the gene trend calculation module is used to determine the trend of each gene based on the TPM value of the transcriptome sequencing gene expression data set of the patient group;

[0086] In this embodiment, the result output module is used to determine the target biomarker and the corresponding reference interval based on the gene trend result.

[0087] Example 3

[0088] This embodiment provides a storage medium, which may be a storage medium such as ROM, RAM, disk, or CD. The storage medium stores one or more programs. When the program is executed by the processor, the method of formulating a biomarker reference interval based on transcriptome sequencing gene expression data of Embodiment 1 is implemented.

[0089] Example 4

[0090] The present embodiment provides a computing device, which may be a desktop computer, a laptop computer, a smart phone, a PDA handheld terminal, a tablet computer or other terminal device with a display function. The computing device includes a processor and a memory, and the memory stores one or more programs. When the processor executes the program stored in the memory, the method of formulating a biomarker reference interval based on transcriptome sequencing gene expression data of Example 1 is implemented.

[0091] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data, characterized in that: The steps include: Obtain transcriptome sequencing data and construct transcriptome sequencing gene expression datasets for healthy and patient groups; Setting screening conditions for low-expression genes, and screening the transcriptome sequencing gene expression data sets of the healthy group and the patient group based on the set screening conditions; Based on the IQR method, outlier samples in the transcriptome sequencing gene expression dataset of the screened healthy group and patient group were detected and removed, and outlier points were filled; Extract the count values ​​of the corresponding genes based on the common gene names in the transcriptome sequencing gene expression datasets of the healthy group and the patient group; Gene length and sequencing depth were standardized, and counts of extracted genes were converted into TPM values; Based on the SMR algorithm, the interval with the most concentrated distribution of TPM values ​​of 95% of the samples in the healthy group transcriptome sequencing gene expression data set was obtained as the reference interval; The trend of each gene was determined based on the TPM value of the transcriptome sequencing gene expression dataset of the patient group; The target biomarkers and corresponding reference intervals were determined based on the gene trend results.

2. The method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data according to claim 1, characterized in that: The transcriptome sequencing gene expression datasets of the healthy group and the patient group were screened based on the set screening conditions, including: For each gene, the proportion of counts values ​​less than the set value is calculated, and the expression values ​​corresponding to genes with a proportion greater than the set proportion are screened out.

3. The method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data according to claim 1, characterized in that: Based on the IQR method, outlier samples in the transcriptome sequencing gene expression dataset of the screened healthy group and patient group were detected and removed, including: For each gene remaining in the healthy group and the patient group after screening, the data outside the interval of [Q1-1.5IQR, Q3+1.5IQR] were screened out, where Q1 was the 25% quantile, Q3 was the 75% quantile, and IQR=Q3-Q1.

4. The method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data according to claim 1, characterized in that: The filling of outliers specifically includes: The average expression of transcriptome sequencing genes after removing outlier samples was calculated and the average value was filled in the outlier points.

5. The method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data according to claim 1, characterized in that: Based on the SMR algorithm, the interval with the most concentrated distribution of TPM values ​​of 95% of the samples in the healthy group transcriptome sequencing gene expression data set was obtained as the reference interval. The calculation formula for the lower and upper limits of the reference interval in the SMR algorithm is expressed as follows: (L i ,U i )={(min(T ik ),max(T ik ))|min(ΔT ik )} ΔT ik =max(T ik )-min(T ik ) Among them, T ik represents the kth subset of the set of TPM values ​​of all samples of the i-th gene in the healthy group, X ij represents the TPM value of the jth sample of the ith gene in the healthy group, j∈[1,n], n represents the total number of samples in the healthy group, and the subset T ik The number of elements k∈[1,a], a is the number of combinations card() indicates the number of elements in the collection. It means that among n samples, 95% of n samples are rounded down and ΔT is taken. ik Represents a subset T ik The difference between the maximum and minimum values, L i Indicates the lower limit of the reference interval, U i Indicates the upper limit of the reference interval; The healthy group was obtained by SMR algorithm The interval where the TPM values ​​of samples are most concentrated [L i , U i ] as the reference interval.

6. The method for formulating biomarker reference intervals based on transcriptome sequencing gene expression data according to claim 1, characterized in that: The trend of each gene was determined based on the TPM value of the transcriptome sequencing gene expression dataset of the patient group, including: Score each sample for the patient group: Among them, Y ih represents the score of the hth sample of the i-th gene in the patient group, S ih represents the TPM value of the hth sample of the ith gene in the patient group; Determine the trend for each gene: Among them, A i is the trend result of gene i, and N represents the total number of samples in the patient group.

7. A system for developing biomarker reference intervals based on transcriptome sequencing gene expression data, characterized in that: include: Data set construction module, gene screening module, outlier data removal module, outlier filling module, gene count value extraction module, TPM value conversion module, reference interval construction module, gene trend calculation module, result output module; The data set construction module is used to obtain transcriptome sequencing data and construct transcriptome sequencing gene expression data sets for the healthy group and the patient group; The gene screening module is used to set screening conditions for low-expression genes, and screen the transcriptome sequencing gene expression data sets of the healthy group and the patient group based on the set screening conditions; The outlier data removal module is used to detect and remove outlier samples in the transcriptome sequencing gene expression data set of the screened healthy group and patient group based on the IQR method; The outlier filling module is used to fill the outliers; The gene counts value extraction module is used to extract the counts values ​​of corresponding genes according to the common gene names in the transcriptome sequencing gene expression data sets of the healthy group and the patient group; The TPM value conversion module is used to standardize the gene length and sequencing depth, and convert the counts value of the extracted gene into a TPM value; The reference interval construction module is used to obtain the interval with the most concentrated distribution of TPM values ​​of 95% of the sample size of the healthy group transcriptome sequencing gene expression data set based on the SMR algorithm as the reference interval; The gene trend calculation module is used to determine the trend of each gene based on the TPM value of the transcriptome sequencing gene expression data set of the patient group; The result output module is used to determine the target biomarker and the corresponding reference interval based on the gene trend result.

8. A computer-readable storage medium storing a program, characterized in that: When the program is executed by a processor, the method for formulating a biomarker reference interval based on transcriptome sequencing gene expression data as described in any one of claims 1 to 6 is implemented.

9. A computer device comprising a processor and a memory for storing a program executable by the processor, characterized in that: When the processor executes the program stored in the memory, the method for formulating a biomarker reference interval based on transcriptome sequencing gene expression data as described in any one of claims 1 to 6 is implemented.

Images