No abstract

Similarity is an important metric in machine learning, which has been applied widely in many fields, such as text matching, [1] image recognition, [2] and biomedical applications. [3][4][5][6] Cosine similarity (CS), Pearson correlation coefficient (PC), and Euclidean distance (ED), as the common similarity measures, have the classical and universal similarity metrics for the analysis of various data. However, these universal methods lose a part of the specific information and hardly utilize the intrinsic characteristics. Each feature of the data has a special meaning and contributes differently to the results, which may be positive or negative. [7,8] Common similarity measures treat all features as the same. A strategy should be considered on how to optimize the feature. Weighting is an efficient approach to quantify the important coefficient of a feature by the data characteristics, which can reduce the influence of weak features and improve the contribution of useful features. Therefore, appropriate weightings can improve the similarity performance to obtain good results in data analysis. [8] Raman spectroscopy based on molecular vibrational scattering is known as the molecular fingerprint. [9,10] The analysis of Raman data is challenging due to the low signal-to-noise ratio, dispersive signals, and signal overlap. [11,12] More and more researchers have explored the analysis of Raman spectra by machine learning. Carey et al. optimized the mineral spectra matching performance using careful preprocessing and a weighting-neighbors classifier of a vector similarity metric. [13] Fenn has presented a novel data analysis framework named fisher-based feature selection support vector machines (FFS-SVM) for classification and has got high accuracy in five cancerous and noncancerous breast cell lines. [14] However, similarity is rarely used for the analysis of Raman spectra by machine learning due to the limited performance of the common similarity.Raman spectroscopy reflects variations of DNA/RNA, proteins, lipid, carbohydrates, and other small-molecule metabolites, which makes it an excellent tool for monitoring biochemical changes on the cellular level. [14][15][16] However, the data analysis methods ignore the biological characteristics of Raman spectra and treat each feature as equally important. Raman spectra are related to the biological characteristics of molecular vibration,

Although the expression of miRNAs has been widely applied to investigate on gonads, the role of miRNAs in the gonadal development of white Pacific shrimp (Litopenaeus vannamei) remains unknown. In this study, we performed high-throughput sequencing to identify the sex-related microRNAs (miRNAs) that elucidated the regulatory mechanisms on the gonadal differentiation of L. vannamei. We obtained a total of 29,671,557 and 28,526,942 raw reads from the ovaries and testes library, respectively. We then mapped 26,365,828 (92.73%) of the ovarian clean sequences and 23,694,294 (85.65%) of the testicular clean sequences for a transcriptome reference sequence of L. vannamei. After blasting the miRNA sequences against the miRBase database, we identified 153 significantly differentially expressed miRNAs between the ovaries and testes. To confirm the high-throughput sequencing results, we used a reverse transcriptase–quantitative polymerase chain reaction (RT-qPCR) to verify the expression patterns of the seven most differentially expressed miRNAs (i.e., novel_mir23, miR-92b-3p_3, miR-12-5p_2, novel_mir67, miR-279_1, let-7-5p_6, miR-263a-5p_1). According to the results of RT-qPCR, most of the miRNAs were expressed consistently with the high-throughput sequencing results. In addition, the target genes significantly enriched several Kyoto Encyclopedia of Genes and Genome (KEGG) pathways that were closely related to gonadal differentiation and development, including extracellular matrix–receptor interaction, Hedgehog signaling pathway, protein digestion and absorption and cell adhesion molecules (CAMs). This study revealed the first miRNAs sequencing of L. vannamei gonads. We identified sex-related differentially expressed miRNAs and KEGG pathways, which will be helpful to facilitate future research into the regulatory mechanism on the gonadal differentiation of L. vannamei.

In the originally published article, the defining of the weighting formula in equations ( 1) and ( 2) and the segment weighting similarity (SWS) in equation ( 3) had been presented incorrectly.