Bidirectional linkage of DNA barcodes for the multiplexed mapping of higher-order protein interactions in cells.

Capturing the full complexity of the diverse hierarchical interactions in the protein interactome is challenging. Here we report a DNA-barcoding method for the multiplexed mapping of pairwise and higher-order protein interactions and their dynamics within cells. The method leverages antibodies conjugated with barcoded DNA strands that can bidirectionally hybridize and covalently link to linearize closely spaced interactions within individual 3D protein complexes, encoding and decoding the protein constituents and the interactions among them. By mapping protein interactions in cancer cells and normal cells, we found that tumour cells exhibit a larger diversity and abundance of protein complexes with higher-order interactions. In biopsies of human breast-cancer tissue, the method accurately identified the cancer subtype and revealed that higher-order protein interactions are associated with cancer aggressiveness.

Deep learning of 2D-Restructured gene expression representations for improved low-sample therapeutic response prediction.

Clinical outcome prediction is important for stratified therapeutics. Machine learning (ML) and deep learning (DL) methods facilitate therapeutic response prediction from transcriptomic profiles of cells and clinical samples. Clinical transcriptomic DL is challenged by the low-sample sizes (34-286 subjects), high-dimensionality (up to 21,653 genes) and unordered nature of clinical transcriptomic data. The established methods rely on ML algorithms at accuracy levels of 0.6-0.8 AUC/ACC values. Low-sample DL algorithms are needed for enhanced prediction capability. Here, an unsupervised manifold-guided algorithm was employed for restructuring transcriptomic data into ordered image-like 2D-representations, followed by efficient DL of these 2D-representations with deep ConvNets. Our DL models significantly outperformed the state-of-the-art (SOTA) ML models on 82% of 17 low-sample benchmark datasets (53% with >0.05 AUC/ACC improvement). They are more robust than the SOTA models in cross-cohort prediction tasks, and in identifying robust biomarkers and response-dependent variational patterns consistent with experimental indications.

Enhanced metagenomic deep learning for disease prediction and consistent signature recognition by restructured microbiome 2D representations.

Metagenomic analysis has been explored for disease diagnosis and biomarker discovery. Low sample sizes, high dimensionality, and sparsity of metagenomic data challenge metagenomic investigations. Here, an unsupervised microbial embedding, grouping, and mapping algorithm (MEGMA) was developed to transform metagenomic data into individualized multichannel microbiome 2D representation by manifold learning and clustering of microbial profiles (e.g., composition, abundance, hierarchy, and taxonomy). These 2D representations enable enhanced disease prediction by established ConvNet-based AggMapNet models, outperforming the commonly used machine learning and deep learning models in metagenomic benchmark datasets. These 2D representations combined with AggMapNet explainable module robustly identified more reliable and replicable disease-prediction microbes (biomarkers). Employing the MEGMA-AggMapNet pipeline for biomarker identification from 5 disease datasets, 84% of the identified biomarkers have been described in over 74 distinct works as important for these diseases. Moreover, the method also discovered highly consistent sets of biomarkers in cross-cohort colorectal cancer (CRC) patients and microbial shifts in different CRC stages.

Toward ordered -omics data science: Researchers on the magic of turning metagenomic chaos into image-like patterns.

Wan Xiang Shen, a postdoctoral researcher at National University of Singapore, and Yu Zong Chen, the PI of the Bioinformatics and Drug Design (BIDD) group, have developed an AI pipeline for enhanced deep learning of metagenomic data. Their Patterns paper highlights the advantages of unsupervised data restructuring in microbiome-based disease prediction and biomarker discovery. They talk about their view of data science and the backstory of the article published in Patterns.

AggMapNet: enhanced and explainable low-sample omics deep learning with feature-aggregated multi-channel networks.

Omics-based biomedical learning frequently relies on data of high-dimensions (up to thousands) and low-sample sizes (dozens to hundreds), which challenges efficient deep learning (DL) algorithms, particularly for low-sample omics investigations. Here, an unsupervised novel feature aggregation tool AggMap was developed to Aggregate and Map omics features into multi-channel 2D spatial-correlated image-like feature maps (Fmaps) based on their intrinsic correlations. AggMap exhibits strong feature reconstruction capabilities on a randomized benchmark dataset, outperforming existing methods. With AggMap multi-channel Fmaps as inputs, newly-developed multi-channel DL AggMapNet models outperformed the state-of-the-art machine learning models on 18 low-sample omics benchmark tasks. AggMapNet exhibited better robustness in learning noisy data and disease classification. The AggMapNet explainable module Simply-explainer identified key metabolites and proteins for COVID-19 detections and severity predictions. The unsupervised AggMap algorithm of good feature restructuring abilities combined with supervised explainable AggMapNet architecture establish a pipeline for enhanced learning and interpretability of low-sample omics data.

[Clinical oberservation on rotator cuff suture threader and strapping suture in meniscus tear under arthroscopy].

OBJECTIVE: To explore clinical effect of arthroscopic meniscus tear strapping suture by rotator cuff suture threader. METHODS: Forty patients with meniscus tear injury admitted from July 2015 to May 2019, including 27 males and 13 females, aged from 20 to 55 years old with an average of (36.0±1.4) years old. Menisci laceration was sutured with rotator cuff suture thread under arthroscopy. Postoperative complication was observed, Lysholm knee joint score before and after operation at 12 months were used to evaluate clinical effects, visual analogue scale (VAS) and range of knee flexion and extension were applied to evaluate recovery of pain and function. RESULTS: All patients were followed up from 12 to 15 months with an average of (12.6±0.7) months.No complication such as joint effusion, suture failure occurred. Two patients occurred mild pain after activity without clinical physical abnormality, and 1 patient manifested moderate pain with joint space tenderness, the other rest without abnormal. Lysholm knee joint score was increased from (49.55±1.21) preoperatively to (98.95±0.42) at 12 months after operation, VAS score decreased from (5.18±0.78)preoperatively to (1.03±0.77) at 12 months after operation, and range of knee joint flexion and extension activity increased from (50.63±9.20)°preoperatively to (130.38±4.99)°after operation, and there were statistical differences in Lysholm knee joint score, VAS and range of knee joint flexion and extension activity (P< 0.05). CONCLUSION: Arthroscopic strapping suture by rotator cuff suture threading device applies to most meniscus injuries, including medial meniscus posterior horn tears, lateral meniscus body tears and lateral meniscus posterior horn tears. This technique meets the need of full-internal meniscus suture without specialmeniscus suture, and has advantages of convenient operation, less complications and good postoperative function.

[Analysis on treatment and diagnosis of Segond fracture combined with injuries].

OBJECTIVE: To investigate the clinical characteristics of Segond fracture combined with injuries, and to explore the operative method and opportunity. METHODS: From June 2010 to December 2014, 10 patients with Segond fracture were treated. There were 10 males, ranging in age from 26 to 69 years old, with a mean of 42 years old. Under arthroscopy examination within 7 to 10 days after injury, 6 patients had compete rupture of anterior cruciate ligament and meniscus injury; 2 patients had tibial avulsion fractures of the ACL combined with meniscus injury. At 4 to 8 weeks after injury, 2 patients were performed with arthroscopy examination; 1 patient had complete rupture of anterior and posterior cruciate ligament combined with meniscus injury; and the other patient had complete rupture of anterior cruciate ligament alone. All the patients with cruciate ligament rupture were treated with reconstruction of cruciate ligament; 9 patients had meniscus injuries were treated with menisus plasty or suture; and 2 patients had tibial avulsion fractures of the ACL were treated with non-absorbable Ethicon suture fixation under arthroscopy. After operation, the Lysholm scoring system was used to evaluate clinical effects. RESULTS: Eight patients were operated within 10 days after injury, the knee joints swelling was obvious at the 3rd day after operation, and blood oozing was found by joint puncture. These 8 patients did out-of-bed activity with assistance at 1 week postoperatively. Other 2 patients performed operation in 4 to 8 weeks after injury without knee joint obvious swelling at the 3rd day after operation, and no blood oozing was found by joint puncture. These 2 patients did out-of-bed activity with assistance after 3 days postoperatively. All the patients were followed up, and the duration ranged from 12 to 24 months, with an average of 18 months. Postoperative Lysholm score of all patients was higher than preoperative scor, and the results were satisfactory. CONCLUSION: For Segond fractures, firstly, it is important to gain early diagnosis for treatment. Secondly, it is often associated with the rupture of anterior cruciate ligament and meniscus injuries, and the best time of cruciate ligament rebuilding and meniscus repairing maybe between 4 to 8 weeks after injury. Thirdly, personalized treatment should be chosen according to different complications.

[Finite element analysis of adolescent idiopathic scoliosis of PUMC II d2 surgical treatment with different fusion levels].

OBJECTIVE: To explore the best surgical fusion level for adolescent idiopathic scoliosis (AIS) of PUMC II d(2) with finite element model (FEM). METHODS: FEM (T(5)-S) of PUMC II d(2) idiopathic scoliosis was used to simulate upper thoracic curve, lower lumbar curve and double curve fusion manners. The pedicle of concave vertebral arch received 50, 100 and 150 N load respectively. Displacement of T(5) and T(11) on upper sagittal plane (displacement of Z axis positive value on upper sagittal plane, displacement of negative value on lower sagittal plane) and their different values were compared. T(5) displacement represented the outcomes of double curve orthopedics. T(11) displacement represented the outcomes of lower lumbar curve orthopedics. Their difference (T(5)-T(11)) represented the outcomes of upper thoracic curve orthopedics. Different fusion segments and displacement of T(5) and T(11) under different orthopedic forces were measured. RESULTS: In PUMC II d(2) lateral curvature, T(5) displacement on Z axis: fusion displacement of double curve was greater than the upper or lower curve alone (F = 8, P < 0.01). Difference of T(5)-T(11) displacement: double curve orthopedics > upper thoracic curve orthopedics alone > lower lumbar curve orthopedics alone (F = 8, P < 0.01). Displacement of T(11) on Z axis: double curve orthopedics > lower lumbar curve orthopedics alone > upper thoracic curve orthopedics alone (F = 8, P < 0.01). CONCLUSION: Fusing two curves achieves the best effect on the AIS of PUMC II d(2) in comparison with upper or lower curve fusion alone. Effects of 3 kinds of load on correction of upper thoracic curve: double curve orthopedics > upper thoracic curve orthopedics alone > lower lumbar curve orthopedics alone. Effect of 3 kinds of load on correction of lower lumbar curve orthopedics alone: double curve orthopedics > lower lumbar curve orthopedics alone > upper thoracic curve orthopedics alone. Three-dimensional finite element analysis is an effective method to analyze the biomechanics of scoliosis deformity correction and provides a virtually non-invasive verification manner. And it may optimize the surgical protocol.

[Reverse fasciocutaneous flap pedicled with perforator branch of anterior medial malleolus artery for soft tissue defect on the dorsal side of foot].

OBJECTIVE: To report the therapeutic effect of reverse fasciocutaneous flap pedicled with perforator branch of anterior medial malleolus artery for soft tissue defect on the dorsal side of foot. METHODS: The perforator branch was located under the guidance of Doppler flowmeter. The flap was designed along the saphenous neurovascular axis. Then the flap was transferred reversely with the perforator branch as rotation point. RESULTS: From Feb. 2002 to Mar. 2008, 12 cases were treated and followed up for 6 - 18 months. All the flaps survived completely. The flap size ranged from 13.5 cm x 3.0 cm to 8 cm x 3 cm. The perforator branch located at 0.5 - 1.5 cm anterior-inferior to the medial malleolus. Both the cosmetic and functional results were satisfactory. CONCLUSIONS: The flap has a reliable blood supply and a flexible design. It is easily performed for soft tissue defect on the dorsal side of foot. It is a new type flap which combined neurocutaneous vascular flap with the perforator flap.